System and method for monitoring progress of thyroid-related eye diseases

Abstract

The invention relates to the technical field of image analysis, in particular to a thyroid-related eye disease progress monitoring system. The progress monitoring method is applied to the eye detector, and comprises the following steps: acquiring an eye monitoring video shot by a camera device, analyzing the eye monitoring video, and determining eye features; analyzing eye features, and determining eyelid opening and closing degrees and eyeball movement conditions; determining pupil change conditions according to the eyeball movement conditions; determining a current eye disease state according to the eyelid opening and closing degree, the eyeball movement condition and the pupil change condition; historical eye disease diagnosis and treatment records are obtained and analyzed, and historical eye disease states are obtained; and comparing the current eye disease state with the historical eye disease state to determine whether the thyroid-related eye disease is improved or not.

Description

Technical Field

[0001] The present application relates to the technical field of image analysis, and in particular to a system and method for monitoring the progression of thyroid-related eye diseases. Background Art

[0002] Thyroid-associated ophthalmopathy is a specific autoimmune disease that affects both the thyroid gland and the orbital organs. It is one of the most common orbital diseases in adults.

[0003] At present, the treatment of thyroid-related eye diseases is mostly through comprehensive analysis through face-to-face consultation combined with CT angiography scans to determine the severity of the patient's thyroid-related eye disease. In the future, if you want to determine the progress of the treatment of thyroid-related eye diseases, you also need to go to the hospital for evaluation by a doctor. This process is not only very troublesome for patients, but also for doctors. Since there is a certain time interval between the initial diagnosis and the re-examination, if there are any problems in the middle that cause the thyroid-related eye disease to become more serious, this is unknown. Therefore, the diagnosis after the re-examination will be the same as the initial diagnosis, which is time-consuming and laborious, and the cause cannot be found, delaying the patient's condition and endangering the patient's health. Summary of the invention

[0004] The present application provides a system and method for monitoring the progression of thyroid-related eye diseases to solve the above-mentioned problems.

[0005] In a first aspect, the present application provides a method for monitoring the progression of thyroid-related eye diseases, the method for monitoring the progression of which is applied to an eye detector, the eye detector comprising a camera device and a data processing chip, the method being applied to the data processing chip; the method comprising: Acquire the eye monitoring video captured by the camera device, analyze the eye monitoring video, and determine eye features; Analyzing the eye characteristics to determine the degree of eyelid opening and closing and eye movement; Determining pupil changes according to the eye movement; Determining the current eye disease state according to the eyelid opening and closing degree, the eyeball movement, and the pupil change; Obtain and analyze historical eye disease diagnosis and treatment records to obtain historical eye disease status; compare the current eye disease status with the historical eye disease status to determine whether the thyroid-related eye disease has improved.

[0006] Through this solution, the eye monitoring video captured by the camera device is obtained in real time to realize the real-time capture of eye status changes. Video analysis and image processing technology provide objective eye feature data, reducing the subjectivity of human judgment. Advanced image processing algorithms are used to accurately measure the degree of eyelid opening and closing, eye movement and pupil changes, improving the accuracy and reliability of diagnosis. Combining multiple eye features for comprehensive judgment avoids the limitations of a single indicator and provides comprehensive information for the assessment of eye disease status. Obtain and analyze historical eye disease diagnosis and treatment records to comprehensively track changes in the patient's condition. By comparing the current eye disease status with the historical eye disease status, the treatment effect can be intuitively evaluated, providing a basis for adjusting the treatment plan. Avoid the situation where the doctor cannot intuitively judge the patient's condition during the follow-up visit due to the long time interval, resulting in the inability to cure the disease in time.

[0007] Optionally, the acquiring the eye monitoring video captured by the camera device includes: Acquire and analyze a starting eye image, and determine whether there is eye deviation based on the image analysis result; If it is determined that there is eye deviation, the eye deviation angle is determined based on the image analysis result; The eye difference is determined according to the eye deviation angle and the historical eye disease diagnosis and treatment records, and the image capture method is adjusted according to the eye difference, and the eye monitoring video is obtained according to the adjusted image capture method.

[0008] Through this solution, the image analysis technology is used to analyze the initial eye image obtained to determine whether the eye is offset, preliminarily determine the eye position and the wearing position of the eye detector, determine the eye offset angle, and then differentiate the image of the eye position based on the eye offset angle and historical eye disease diagnosis and treatment records to prepare for the subsequent image adjustment. By determining the image capture method and adjusting the shooting area, it is ensured that even if the patient's wearing is not standard, accurate capture can be performed, reducing video errors and improving the accuracy of subsequent eye disease status judgment.

[0009] Optionally, the eye detector includes a camera device; the acquiring and analyzing the initial eye image, and determining whether there is eye deviation according to the image analysis result, includes: monitoring the displacement change of the eye detector in real time, and determining whether the spatial position of the eye detector changes according to the displacement change; If it is determined that there is a spatial position change, determining a horizontal position change amplitude according to the spatial position change; Determining whether to start the camera device according to the horizontal position change amplitude; If it is determined to start the camera device, acquiring and screening the captured image of the camera device to obtain the initial eye image; Analyze the initial eye image, and extract eye features according to the image analysis result; According to the eye features, a structure boundary is determined, and according to the structure boundary, the eye features in each structure are analyzed to determine whether there is an eye offset.

[0010] Through this solution, by integrating high-precision sensors and real-time data processing technology, the displacement changes of the eye detector can be monitored in real time. When the spatial position of the eye detector changes, it is immediately decided whether to start the camera device to shoot according to the amplitude of the spatial position change, and the automatic camera device can be started without manual intervention. The images taken by the camera device are screened to ensure that a clear and accurate starting eye image is obtained. Subsequently, the eye features are extracted using advanced image recognition technology to further improve the accuracy of the analysis. After extracting the eye features, the structural boundaries of the eye are determined based on these features. Structural analysis helps to more accurately locate and analyze the eye displacement phenomenon.

[0011] Optionally, analyzing the eye features in each structure according to the structure boundary to determine whether eye offset exists includes: determining an eyelid region based on the structure boundaries and eye features within each structure; Analyzing eye features of the eyelid region to determine an eyelid shadow area; Determining the eyelid shadow area according to the eyelid shadow area; Obtaining current lighting conditions and the spatial position of the camera device; Determining a desired eyelid shadow area according to the current lighting condition and the spatial position; Comparing the eyelid shadow area with the expected eyelid shadow area to determine whether the eyelid shadow area is greater than the expected eyelid shadow area; If the eyelid shadow area is greater than the expected eyelid shadow area, it is determined that eye deviation exists.

[0012] Through this solution, the eye features within each structure are carefully analyzed and the eyelid area is precisely located, which helps to more accurately capture the changes in eye features. This solution not only focuses on the eye features themselves, but also takes into account the current lighting conditions and the spatial position of the camera device. By incorporating these environmental factors into the analysis, the expected eyelid shadow area is more comprehensively evaluated. The actual eyelid shadow area is compared with the expected eyelid shadow area to intelligently determine whether there is an eye offset. This automated judgment method reduces the errors and uncertainties caused by human factors.

[0013] Optionally, analyzing the eye monitoring video to determine eye features includes: Analyzing the eye monitoring video to determine changes in light and dark; Determining a current detection mode according to the light and dark changes; According to the current detection mode, the eye monitoring video is analyzed frame by frame to determine eye features.

[0014] This solution analyzes eye monitoring videos and determines the detection mode based on changes in light and dark. Through refined frame-by-frame analysis and dynamically adjusted detection modes, eye features can be more effectively identified, thereby reducing false positives and false negatives. This helps improve overall performance and user experience.

[0015] Optionally, analyzing the eye features to determine the degree of eyelid opening and closing and eyeball movement includes: Analyzing the eye features to determine the eyelid contour and eyeball area; Determining the relative position between the eyelid contour and the eyeball area according to the feature vector of the eye feature; Determine the opening and closing ratio of the eyelid according to the relative position, and determine the opening and closing degree of the eyelid according to the opening and closing ratio; Determine the pupil position in each frame of video according to the eye features; According to the pupil position in each frame of video, the pupil change speed, pupil change direction and pupil change acceleration are determined; The eye movement condition is determined according to the pupil change speed, the pupil change direction and the pupil change acceleration.

[0016] Through this solution, the relative position between the eyelid contour and the eyeball area is calculated, the eyelid opening and closing ratio is accurately determined, and then the degree of eyelid opening and closing is accurately judged. The pupil position in each frame of the video is tracked in real time, which makes it possible to calculate the pupil change speed, direction and acceleration. Taking into account the pupil change speed, direction and acceleration, a comprehensive evaluation of the eyeball's zooming can be achieved, and a more accurate eyeball movement can be obtained, which is more convenient for the subsequent detection and diagnosis of patients' eye diseases.

[0017] Optionally, determining pupil changes according to the eye movement includes: Based on the eye movement, real-time acquisition of environmental changes; determining a pupil inspection mode according to the environmental changes; Determining the real-time position of the pupil according to the eye movement; Determining a real-time pupil diameter based on the real-time pupil position and the eye feature; The pupil change is determined according to the pupil inspection mode and the real-time pupil diameter.

[0018] Through this solution, through real-time monitoring of eye movement, the light of the surrounding environment when the eye is zoomed in and out can be quickly captured, and the pupil examination mode can be determined based on the determined changes in the surrounding environment, so that the analysis of pupil changes can be closely linked to different examination modes, which is convenient for subsequent analysis of the patient's condition. In addition, it is more in line with the patient's eye examination needs. Changes in pupil diameter can directly reflect the degree of pupil response to different environmental stimuli, which helps to evaluate the patient's visual adjustment ability, making it easier to analyze the patient's condition.

[0019] Optionally, the eye movement condition is determined according to the pupil change speed, the pupil change direction and the pupil change acceleration, referring to the following formula: ; in, represents the eye movement index; represents the pupil change speed; A weight coefficient representing the pupil change speed; Indicates the average value of the pupil change speed within a preset detection period; represents the standard deviation of the pupil change speed; A weight coefficient representing the direction of pupil change; Indicates the quantified direction of pupil change; A weight coefficient representing the acceleration of pupil change; represents the pupil change acceleration; represents the acceleration mean of the pupil change acceleration within a preset detection period; represents the standard deviation of the pupil change acceleration; The eye movement condition is determined according to the eye movement index.

[0020] Through this solution, the pupil change speed, pupil change direction and pupil change acceleration are taken into consideration, and the eyeball scaling index is first obtained by designing a formula calculation, so as to use the eyeball scaling index to reflect the eyeball movement. On the one hand, the evaluation of the eyeball movement can be made more comprehensive and accurate. On the other hand, the design formula can digitize complex content, so as to more simply reflect the eyeball movement.

[0021] Optionally, determining the current eye disease state according to the eyelid opening and closing degree, the eyeball movement, and the pupil change includes: Determining the last inspection data according to the current inspection mode; Determining the last eyelid opening and closing degree, the last eyeball zooming degree and the last pupil change according to the last examination data; Determining the eye change trend according to the eyelid opening and closing degree, the eyeball movement condition, the pupil change condition, the last eyelid opening and closing degree, the last eyeball zooming degree and the last pupil change condition; The current state of the eye disease is determined based on the eye change trend.

[0022] This solution integrates data from multiple dimensions, such as eyelid opening and closing, eye movement, pupil changes, etc., to provide a solid foundation for a comprehensive assessment of eye conditions. By comparing the current test data with the last test data, the changing trend of eye conditions can be accurately captured. By analyzing the trend of eye changes, the current state of eye disease can be determined, providing a basis for subsequent judgment of whether the eye disease has improved.

[0023] In a second aspect, the present application provides a system for monitoring the progression of thyroid-related eye diseases, the system comprising: A feature determination module, used to obtain an eye monitoring video taken by a camera device, analyze the eye monitoring video, and determine eye features; A feature analysis module, used to analyze the eye features and determine the degree of eyelid opening and closing and eyeball movement; A pupil analysis module, used to determine pupil changes according to the eye movement; A state analysis module, which determines the current state of eye disease according to the degree of eyelid opening and closing, the eyeball movement, and the pupil change; The symptom analysis module is used to obtain and analyze historical eye disease diagnosis and treatment records to obtain historical eye disease status; compare the current eye disease status with the historical eye disease status to determine whether the thyroid-related eye disease has improved.

[0024] Optionally, the feature determination module is specifically used to: Acquire and analyze a starting eye image, and determine whether there is eye deviation based on the image analysis result; If it is determined that there is eye deviation, the eye deviation angle is determined based on the image analysis result; The eye difference is determined according to the eye deviation angle and the historical eye disease diagnosis and treatment records, and the image capture method is adjusted according to the eye difference, and the eye monitoring video is obtained according to the adjusted image capture method.

[0025] Optionally, the eye detector includes a camera device; and the feature determination module is specifically used to: monitoring the displacement change of the eye detector in real time, and determining whether the spatial position of the eye detector changes according to the displacement change; If it is determined that there is a spatial position change, determining a horizontal position change amplitude according to the spatial position change; Determining whether to start the camera device according to the horizontal position change amplitude; If it is determined to start the camera device, acquiring and screening the captured image of the camera device to obtain the initial eye image; Analyze the initial eye image, and extract eye features according to the image analysis result; According to the eye features, a structure boundary is determined, and according to the structure boundary, the eye features in each structure are analyzed to determine whether there is an eye offset.

[0026] Optionally, the feature determination module is specifically used to: determining an eyelid region based on the structure boundaries and eye features within each structure; Analyzing eye features of the eyelid region to determine an eyelid shadow area; Determining the eyelid shadow area according to the eyelid shadow area; Obtaining current lighting conditions and the spatial position of the camera device; Determining a desired eyelid shadow area according to the current lighting condition and the spatial position; Comparing the eyelid shadow area with the expected eyelid shadow area to determine whether the eyelid shadow area is greater than the expected eyelid shadow area; If the eyelid shadow area is greater than the expected eyelid shadow area, it is determined that eye deviation exists.

[0027] Optionally, the feature determination module is specifically used to: Analyzing the eye monitoring video to determine changes in light and dark; Determining a current detection mode according to the light and dark changes; According to the current detection mode, the eye monitoring video is analyzed frame by frame to determine eye features.

[0028] Optionally, the feature analysis module is specifically used to: Analyzing the eye features to determine the eyelid contour and eyeball area; Determining the relative position between the eyelid contour and the eyeball area according to the feature vector of the eye feature; Determine the opening and closing ratio of the eyelid according to the relative position, and determine the opening and closing degree of the eyelid according to the opening and closing ratio; Determine the pupil position in each frame of video according to the eye features; According to the pupil position in each frame of video, the pupil change speed, pupil change direction and pupil change acceleration are determined; The eye movement condition is determined according to the pupil change speed, the pupil change direction and the pupil change acceleration.

[0029] Optionally, the pupil analysis module is specifically used for: Based on the eye movement, real-time acquisition of environmental changes; determining a pupil inspection mode according to the environmental changes; Determining the real-time position of the pupil according to the eye movement; Determining a real-time pupil diameter based on the real-time pupil position and the eye feature; The pupil change is determined according to the pupil inspection mode and the real-time pupil diameter.

[0030] Optionally, the feature analysis module is specifically used to: ; in, represents the eye movement index; represents the pupil change speed; A weight coefficient representing the pupil change speed; Indicates the average value of the pupil change speed within a preset detection period; represents the standard deviation of the pupil change speed; A weight coefficient representing the direction of pupil change; Indicates the quantified direction of pupil change; A weight coefficient representing the acceleration of pupil change; represents the pupil change acceleration; represents the acceleration mean of the pupil change acceleration within a preset detection period; represents the standard deviation of the pupil change acceleration; The eye movement condition is determined according to the eye movement index.

[0031] Optionally, when determining the current eye disease state according to the eyelid opening and closing degree, the eyeball movement, and the pupil change, the state analysis module is specifically used to: Determining the last inspection data according to the current inspection mode; Determining the last eyelid opening and closing degree, the last eyeball zooming degree and the last pupil change according to the last examination data; Determining the eye change trend according to the eyelid opening and closing degree, the eyeball movement condition, the pupil change condition, the last eyelid opening and closing degree, the last eyeball zooming degree and the last pupil change condition; The current state of the eye disease is determined based on the eye change trend. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0033] Figure 1 A schematic diagram of an application scenario provided for an embodiment of the present application; Figure 2 A flow chart of a method for monitoring the progression of thyroid-related eye disease provided in one embodiment of the present application; Figure 3 A schematic diagram of the structure of a system for monitoring the progression of thyroid-related eye diseases provided in one embodiment of the present application. DETAILED DESCRIPTION

[0034] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0035] In addition, the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article, unless otherwise specified, generally means that the associated objects before and after are in an "or" relationship.

[0036] The embodiments of the present application are further described in detail below in conjunction with the drawings in the specification.

[0037] If you want to determine the progress of treatment for thyroid-related eye disease in the future, you need to go to the hospital for evaluation by a doctor. This process is not only very troublesome for patients, but also for doctors. There is a certain time interval between the initial diagnosis and the re-examination. If there are any problems in the middle that cause the thyroid-related eye disease to become more serious, this is unknown. Therefore, the diagnosis after the re-examination will be the same as the initial diagnosis, which is time-consuming and laborious, and the cause cannot be found, which delays the patient's condition and endangers the patient's health.

[0038] Based on this, the present application provides a system and method for monitoring the progression of thyroid-related eye diseases, which can obtain eye monitoring videos taken by a camera in real time and realize real-time capture of changes in eye status. Video analysis and image processing technology provide objective eye feature data, reducing the subjectivity of human judgment. Advanced image processing algorithms are used to accurately measure the degree of eyelid opening and closing, eye movement, and pupil changes, thereby improving the accuracy and reliability of diagnosis. A comprehensive judgment is made by combining multiple eye features to avoid the limitations of a single indicator and provide comprehensive information for the evaluation of eye disease status. Historical eye disease diagnosis and treatment records are obtained and analyzed to comprehensively track changes in the patient's condition. By comparing the current eye disease status with the historical eye disease status, the treatment effect can be intuitively evaluated, providing a basis for adjusting the treatment plan. It avoids the situation where the doctor cannot intuitively judge the patient's condition during follow-up visits due to long time intervals, resulting in the inability to cure the disease in time.

[0039] Figure 1 This is a schematic diagram of an application scenario provided in the present application. When it is necessary to monitor the progress of treatment of thyroid-related eye diseases, the method provided in the present application can be applied.

[0040] Specifically, the method provided in the present application is applied to any eye detector, and the eye detector interacts with the server of the attending physician, and a camera device and a data processing chip are provided in the eye detector. The eye monitoring video taken by the camera device is obtained, and transmitted to the data processing chip for analysis of the eye monitoring video, and the eye features are determined and analyzed, the degree of eyelid opening and closing and the eyeball movement are determined, and then the eyeball movement is analyzed, and the pupil changes are determined. The current state of the eye disease is determined by combining the degree of eyelid opening and closing, eyeball movement, and pupil changes; the patient's historical eye disease diagnosis and treatment records in the attending physician's server are retrieved and analyzed to obtain the historical eye disease state, and the current eye disease state is compared with the historical eye disease state to determine whether there is improvement in thyroid-related eye disease. Video analysis and image processing technology provide objective eye feature data, reducing the subjectivity of human judgment. Combining multiple eye features for comprehensive judgment avoids the limitations of a single indicator and provides comprehensive information for the evaluation of eye disease status.

[0041] For specific implementation methods, please refer to the following embodiments.

[0042] Figure 2 This is a flow chart of a method for monitoring the progression of thyroid-related eye disease provided in one embodiment of the present application. The method of this embodiment can be applied to the eye detector in the above scenario. Figure 2 As shown, the method includes: S201: Obtain an eye monitoring video taken by a camera device, analyze the eye monitoring video, and determine eye features.

[0043] The eye detector can be understood as an instrument worn in the eye area, with a physical start button. Patients who use the eye detector for eye testing can choose at what point to start, such as after wearing is completed. When the physical button is pressed to send a start signal, the camera device is automatically turned on to shoot. All videos generated during the shooting process can be filtered to eliminate useless frames, such as images of non-eye areas taken by patients when they start before wearing is completed. After the screening is completed, the most popular eye monitoring video is obtained, and the eye monitoring video is analyzed using image analysis technology to determine the eye area and extract eye features.

[0044] In some implementations, strain sensors and gyroscopes can be installed in the eye detector's straps. The strain sensors sense the stretching or compression of the straps to determine the tightness of the straps. Generally, when a patient changes from never wearing the straps to wearing the straps, the straps will change from the original state to the stretched state. At this time, the strain sensor receives the data of the strap stretching, which can be considered as the patient wearing the straps. At this time, in conjunction with the gyroscope in the eye detector, it senses whether the eye detector is in a stable state, that is, it changes from a shaking state to a static state, which indicates the process from someone picking it up to wearing it. If it is in a stable state, a start signal can be sent to the camera device to collect video.

[0045] S202: Analyze eye features to determine the degree of eyelid opening and closing and eyeball movement.

[0046] The eye features obtained above are analyzed to determine the eyelid area and the eyeball area, and then the eye features are distinguished according to the frame images. By analyzing the pixel changes of the eye features in each frame image, the changes in the eyelid area features are determined, and the distance from the eyelid edge to the eyeball center in each frame image is calculated to obtain the degree of eyelid opening and closing. In a specific implementation, indicators such as the eye aspect ratio can also be used to determine the degree of eyelid opening and closing.

[0047] Similarly, after distinguishing the eye features in each frame in the above manner, the eye movement is determined by tracking the position and range of the pixel points in the eye area of ​​each frame.

[0048] S203: Determine pupil changes according to eye movement.

[0049] After obtaining the eye movement situation, the eye area in each frame is further distinguished by feature pixels to determine the eye pixels and pupil pixels. The size of the exit pupil is determined through the pupil pixels, and then the changes in the size of each frame are analyzed frame by frame to determine the changes in the exit pupil.

[0050] S204: Determine the current eye disease status according to the degree of eyelid opening and closing, eyeball movement, and pupil changes.

[0051] A database is pre-set in the data processing chip of the eye detector, which is used to store the normal eye conditions corresponding to different age groups and different eye features, as well as the basic information of the patient using the eye detector, such as age, gender, etc. After the eyelid opening and closing degree, eye movement, and pupil change are obtained in the above manner, first, according to the patient's basic information, the corresponding age group and the patient's eye features, such as single eyelids, eye distance, eye length, etc. are determined, and according to these features, the corresponding normal eye condition is retrieved from the preset database, and then the corresponding eyelid opening and closing degree, eye movement, and pupil change under the normal eye state are compared with the current eyelid opening and closing degree, eye movement, and pupil change, so as to determine whether there is an eye disease and the current eye disease state.

[0052] In a specific implementation method, the preset database may also store eye features of several known eye diseases at different degrees. When it is determined through the above steps that there is a difference from the normal eye state, the current eyelid opening and closing degree, eye movement, and pupil changes can be used to select data that matches the state from the eye features of known eye diseases at different degrees, and the eye disease and degree of disease corresponding to this state can be used as the patient's current eye disease state.

[0053] S205. Obtain and analyze historical eye disease diagnosis and treatment records to obtain historical eye disease status; compare the current eye disease status with the historical eye disease status to determine whether there is improvement in thyroid-related eye disease.

[0054] The patient's historical eye disease diagnosis and treatment records are retrieved through the server of the attending physician of China Unicom. In a specific implementation method, when the attending physician diagnoses the patient and confirms that the patient needs to use an eye detector for continuous eye disease progression monitoring, the diagnosis record is directly transmitted to the corresponding eye detector. These diagnosis records can be stored in the above-mentioned preset database as historical eye disease diagnosis and treatment records, and the historical eye disease diagnosis and treatment records can be retrieved directly from the preset database.

[0055] The obtained historical eye disease diagnosis and treatment records are subjected to natural language processing to determine the historical eye disease status described in the records. The historical eye disease status is compared with the current eye disease status to determine whether it has improved. For example, a larger eyelid opening and closing range indicates improvement.

[0056] Through the scheme provided by this embodiment, the eye monitoring video taken by the camera device is obtained in real time, and the real-time capture of eye state changes is achieved. Video analysis and image processing technology provide objective eye feature data, reducing the subjectivity of human judgment. Using advanced image processing algorithms, the degree of eyelid opening and closing, eye movement and pupil changes are accurately measured, improving the accuracy and reliability of diagnosis. Combining multiple eye features for comprehensive judgment avoids the limitations of a single indicator and provides comprehensive information for the assessment of eye disease status. Obtain and analyze historical eye disease diagnosis and treatment records to comprehensively track changes in the patient's condition. By comparing the current eye disease status with the historical eye disease status, the treatment effect can be intuitively evaluated, providing a basis for adjusting the treatment plan. Avoid the situation where the doctor cannot intuitively judge the patient's condition during the follow-up visit due to the long time interval, resulting in the inability to cure the disease in time.

[0057] In some embodiments, an initial eye image is acquired and analyzed, and based on the image analysis results, it is determined whether eye deviation exists; if it is determined that eye deviation exists, the eye deviation angle is determined based on the image analysis results; eye differences are determined based on the eye deviation angle and historical eye disease diagnosis and treatment records, and based on the eye differences, the screen capture method is adjusted, and based on the adjusted screen capture method, the eye monitoring video taken by the camera device is acquired.

[0058] The initial eye image can be considered as the first frame captured after the camera device is started.

[0059] The image capture method can be considered to utilize the self-correction image technology within the device to adjust the layout of the offset image so that the video can be captured accurately without adjusting the eye detector.

[0060] Specifically, after starting the camera device in the above manner, the camera device can be set to preferentially take an image, namely, a starting eye image, to analyze the starting eye image and determine whether there is a positional offset when the patient wears the device. At this time, the position of the eye in the image corresponding to a wear without positional offset can be determined based on the historical wearing records, and then, if the eye is not completely in this position without correction, it can be considered that there is an eye offset. At the same time, when it is determined that there is an offset, a plane rectangular coordinate system is established based on the relationship between the two corresponding positions, and the angle between the two is determined to obtain the eye offset angle. In a specific implementation method, the position of the eye in the image corresponding to a wear without positional offset can be obtained by the attending physician wearing the device before the first wear. This position can be stored in the historical eye disease diagnosis and treatment records.

[0061] When there is eye offset, the eye position in the initial eye image can be compared with the corresponding eye position in the image when the wearer has no position offset, so as to determine the eye difference, and perform automatic correction based on this eye difference. After the correction, it is determined that the eye position in the initial eye image is consistent with the corresponding eye position in the image when the wearer has no position offset, and then the eye monitoring video can be shot and acquired.

[0062] In a specific implementation, if it is not the first time for the patient to wear the eye detector, that is, the patient has used it multiple times and has usage data, the eye difference can be determined based on the angle when it was last worn.

[0063] Through the scheme provided in this embodiment, the acquired initial eye image is analyzed by image analysis technology to determine whether the eye is offset, preliminarily determine the eye position and the wearing position of the eye detector, determine the eye offset angle, and then differentiate the image of the eye position based on the eye offset angle and historical eye disease diagnosis and treatment records to prepare for the subsequent image adjustment. By determining the image capture method and adjusting the shooting area, it is ensured that even when the patient's wearing is not standard, accurate capture can be performed, reducing video errors and improving the accuracy of subsequent eye disease status judgment.

[0064] In some embodiments, the displacement changes of the eye detector are monitored in real time, and based on the displacement changes, it is determined whether there is a spatial position change of the eye detector; if it is determined that there is a spatial position change, then the amplitude of the horizontal position change is determined based on the spatial position change; based on the amplitude of the horizontal position change, it is determined whether to start the camera device; if it is determined that the camera device is started, the shooting picture of the camera device is acquired and screened to obtain an initial eye image; the initial eye image is analyzed, and eye features are extracted based on the image analysis results; based on the eye features, the structural boundaries are determined, and based on the structural boundaries, the eye features within each structure are analyzed to determine whether there is eye offset.

[0065] A high-precision gyroscope can be integrated in the eye detector to analyze the position changes of the eye detector in real time, obtain the displacement changes of the eye detector, and determine whether there is a spatial position change of the eye detector. At this time, someone may want to pick it up for observation rather than for an eye examination. Therefore, it is necessary to determine the amplitude of the horizontal position change through the spatial position change. For example, a spatial rectangular coordinate system is established based on the position of the eye detector before the change, and then the change of the spatial coordinates is determined based on the continuous spatial position change, so as to obtain the final horizontal position change amplitude based on the change of the spatial coordinates.

[0066] When a certain horizontal position change amplitude is reached, the change amplitude can be further analyzed, such as whether there is ups and downs in position, or other position changes. If the horizontal position rises straight up and stabilizes at a certain moment, it can be considered that an eye examination may be needed at this time. When this happens, it can be determined to automatically start the camera device.

[0067] After starting the camera device, it can continuously capture images. At this time, the captured images can be screened to obtain one or more complete eye images related to the eyes, and these complete eye images can be further screened to obtain the complete eye image with the highest clarity as the starting eye image.

[0068] After obtaining the initial eye image, the image analysis technology is used to extract eye features, so as to determine the structural boundaries of the eye according to the spatial distribution and morphological characteristics of these features, and perform a more detailed eye feature analysis to determine whether there is eye deviation.

[0069] Through the solution provided by this embodiment, by integrating high-precision sensors and real-time data processing technology, the displacement changes of the eye detector can be monitored in real time. When the spatial position of the eye detector changes, it is immediately decided whether to start the camera device for shooting according to the amplitude of the spatial position change, and the automatic camera device can be started without manual intervention. The pictures taken by the camera device are screened to ensure that a clear and accurate starting eye image is obtained. Subsequently, the eye features are extracted using advanced image recognition technology to further improve the accuracy of the analysis. After extracting the eye features, the structural boundaries of the various eyes are determined based on these features. Structured analysis helps to more accurately locate and analyze the eye displacement phenomenon.

[0070] In some embodiments, the eyelid region is determined based on the structure boundary and the eye features within each structure; the eye features of the eyelid region are analyzed to determine the eyelid shadow region; the eyelid shadow area is determined based on the eyelid shadow region; the current lighting conditions and the spatial position of the camera device are obtained; the expected eyelid shadow area is determined based on the current lighting conditions and the spatial position; the eyelid shadow area is compared with the expected eyelid shadow area to determine whether the eyelid shadow area is larger than the expected eyelid shadow area; if the eyelid shadow area is larger than the expected eyelid shadow area, it is determined that there is an eye offset.

[0071] The expected eyelid shadow area can be the area of ​​the eyelid shadow when the patient is undergoing an eye examination using the above-mentioned eye detector and is wearing the device correctly and under consistent environmental conditions. It can be determined based on the image obtained by the attending physician when wearing the device in the above-mentioned embodiment.

[0072] By using the structure boundaries obtained in the above embodiment and the eye features in each structure, it is determined what each structure belongs to, and from there, which part belongs to the eyelid region. The eye features of the eyelid region are analyzed, including the position of the eyelid and the shape features of the eyelid, so as to calculate the shadow area of ​​the eyelid region by using the pixel counting method.

[0073] By building a light sensor into the eye detector, the current light conditions when the initial eye image is taken are obtained, and then the eyelid occlusion and the expected eyelid shadow area when the eyelid is in normal wear when the initial eye image is taken are determined in combination with the spatial position of the camera device. This can be predicted by deep learning in combination with the patient's first normal wear.

[0074] In a specific implementation, image processing technology may also be used to determine the illumination intensity and illumination direction of the initial eye image.

[0075] After the expected eyelid shadow area is determined, it is compared with the actual eyelid shadow area. If the eyelid shadow area is larger than the expected eyelid shadow area, it is determined that there is eye deviation.

[0076] In a specific implementation, an area threshold may also be set. If the eyelid shadow area is larger than the expected eyelid shadow area, and the difference between the two is larger than the area threshold, it can be considered that eye deviation exists.

[0077] Through the solution provided by this embodiment, the eye features in each structure are carefully analyzed, and the eyelid area is accurately located, which helps to capture the changes in eye features more accurately. This solution not only focuses on the eye features themselves, but also takes into account the current lighting conditions and the spatial position of the camera device. By incorporating these environmental factors into the analysis, the expected eyelid shadow area is more comprehensively evaluated. The actual eyelid shadow area is compared with the expected eyelid shadow area to intelligently determine whether there is an eye offset. This automated judgment method reduces the errors and uncertainties caused by human factors.

[0078] In some embodiments, the eye monitoring video is analyzed to determine changes in light and dark; based on the changes in light and dark, a current detection mode is determined; based on the current detection mode, the eye monitoring video is analyzed frame by frame to determine eye features.

[0079] The current detection mode may be the eye detection mode currently being performed by the patient. In order to achieve a more comprehensive detection, several detection modes may be preset. Different detection modes have different characteristics. For example, when performing pupil detection, obvious alternation of light and dark may occur.

[0080] Specifically, after obtaining the eye monitoring video, each frame of the eye monitoring video is calculated to determine the global average brightness or histogram distribution of each frame to evaluate the overall brightness change. According to the overall brightness range of the video, a threshold for brightness change is set. When the brightness difference between adjacent frames exceeds the threshold, it is considered that a significant brightness change has occurred. The current detection mode is determined based on the possible sensitive changes in several pre-set detection modes and the brightness changes in each current frame. After determining the current detection mode, a frame-by-frame analysis is performed based on the characteristics of the current detection mode to determine the characteristics of the eye under the current detection mode.

[0081] Through the solution provided in this embodiment, the eye monitoring video is analyzed, and the detection mode is determined in combination with the light and dark changes. Through refined frame-by-frame analysis and dynamically adjusted detection modes, eye features can be more effectively identified, thereby reducing false positives and missed positives. This helps to improve overall performance and user experience.

[0082] In some embodiments, eye features are analyzed to determine the eyelid contour and eyeball area; based on the feature vector of the eye features, the relative position between the eyelid contour and the eyeball area is determined; based on the relative position, the opening and closing ratio of the eyelid is determined, and based on the opening and closing ratio, the degree of eyelid opening and closing is determined; based on the eye features, the pupil position in each frame of video is determined; based on the pupil position in each frame of video, the pupil change speed, pupil change direction and pupil change acceleration are determined; based on the pupil change speed, pupil change direction and pupil change acceleration, the eyeball movement is determined.

[0083] The eyelid opening and closing ratio can be understood as the ratio of the eyelid open part to the fully open state when the eyes are in the open state.

[0084] The eyelid opening degree can be understood as the current eyelid opening degree compared to the eyelid opening degree when the patient is in a normal eye state.

[0085] Specifically, the eye features obtained in the above embodiment are first processed to reduce the occurrence of errors. The area excluding the eyelid is identified using an edge detection algorithm, and the outline of the eyelid is clarified. Then, all eye features are classified using a color segmentation method, and texture analysis is performed on the classified eye features to determine the relevant features of the eyeball, and the area excluding the eyeball is determined based on the color and texture of these features.

[0086] Based on the eyelid contour and eyeball area determined above, feature vector conversion is performed on the eye features at the corresponding positions, the feature vector of the eyelid contour and the feature vector of the eyeball area are compared, and the relative position relationship between the two, such as distance, angle, etc., is calculated.

[0087] In order to better judge the degree of eyelid opening and closing, an opening and closing ratio threshold can be set in advance according to the patient's eye condition. This threshold can be set multiple times to reflect the degree of eyelid opening and closing under different eyelid opening and closing ratios, such as fully open, half open, closed, etc.

[0088] According to the eye features corresponding to the eye area, using the grayscale, shape and other features of the pupil, machine learning is used to locate the position of the pupil, or according to the changes of the feature vector corresponding to the eye features in each frame of video, the pupil position in each frame of video is determined; based on the pupil position in each frame of video and combined with the speed of change of each frame, the pupil change speed, pupil change direction and pupil change acceleration are determined; according to the pupil change speed, pupil change direction and pupil change acceleration, the eye movement is determined.

[0089] The pupil change speed can be obtained by calculating the moving distance of the pupil in the continuous frames and dividing it by the time interval corresponding to the change speed of each frame. Assuming that the pupil is The frame position is , in The frame position is , the time interval is , then the pupil change speed It can be calculated by referring to the following formula (1): (1); In addition, the direction of pupil change can be determined by the change trend of pupil position. This can include moving toward the center and spreading outward, quantifying the change trend of pupil position. Here, a threshold is assumed. , when it is less than this threshold, it means that the pupil is moving toward the center. In other cases, it can be considered that the pupil is spreading outward. For details, please refer to the following formula (2) for calculation: (2); The acceleration of pupil change can be calculated by the second-order derivative or differential method to calculate the acceleration of pupil position change. For details, please refer to the following formula (3): (3); Through the scheme provided by this embodiment, the relative position between the eyelid contour and the eyeball area is calculated, the opening and closing ratio of the eyelid is accurately determined, and then the opening and closing degree of the eyelid is accurately judged, and the pupil position in each frame of the video is tracked in real time, which makes it possible to calculate the pupil change speed, direction and acceleration. By comprehensively considering the pupil change speed, direction and acceleration, a comprehensive evaluation of the zooming of the eyeball is achieved, and a more accurate eyeball movement is obtained, which is more convenient for the subsequent detection and diagnosis of the patient's eye disease.

[0090] In some embodiments, environmental changes are acquired in real time based on eye movements; a pupil inspection mode is determined based on environmental changes; the real-time position of the pupil is determined based on eye movements; the real-time pupil diameter is determined based on the real-time pupil position and eye features; pupil changes are determined based on the pupil inspection mode and the real-time pupil diameter.

[0091] Based on the eyeball movement conditions obtained above, each time the eyeball changes, the current environmental changes, such as the lighting conditions, are captured in real time, and the current pupil inspection mode is determined based on the environmental changes and the several detection modes set in the above embodiments. Then, based on the above eyeball movement conditions, the real-time position of the pupil at each moment is determined, that is, the position of the pupil in the eye at the current moment.

[0092] Thus, the pupil diameter at the current moment, i.e., the real-time pupil diameter, is determined by combining the pupil real-time position and the eye features at the current moment. The pupil change is determined according to the pupil inspection mode and the real-time pupil diameter.

[0093] In a specific implementation, the patient's condition can be used to establish a correlation between eye zoom characteristics and environmental factors, and then a model can be established to determine environmental changes under different eye zoom conditions. In other implementations, the current image can also be analyzed to determine the distribution of light in the image, and then combined with the sensors set inside the eye detector, the light intensity and light distribution of the patient's environment can be determined, thereby inferring environmental changes.

[0094] Through the solution provided by this embodiment, the light of the surrounding environment when the eye is zoomed is quickly captured through real-time monitoring of eye movement, and the pupil examination mode is determined based on the determined changes in the surrounding environment, so that the analysis of pupil changes can be closely linked with different examination modes, which is convenient for subsequent analysis of the patient's condition. In addition, it is more in line with the patient's eye examination needs. The change in pupil diameter can directly reflect the degree of pupil response to different environmental stimuli, which helps to evaluate the patient's visual adjustment ability, thereby making it easier to analyze the patient's condition.

[0095] In some embodiments, an eye movement index is determined based on the pupil change speed, the pupil change direction and the pupil change acceleration, and the eye movement condition is determined based on the eye movement index.

[0096] According to the pupil change speed, pupil change direction and pupil change acceleration, the eye movement is determined, referring to the following formula (4): (4); in, represents the eye movement index; Indicates the speed of pupil change; The weight coefficient representing the pupil change speed; Indicates the average value of pupil change speed within the preset detection period; represents the standard deviation of pupil change speed; The weight coefficient representing the direction of pupil change; Indicates the direction of pupil change after quantification; The weight coefficient representing the acceleration of pupil change; represents the acceleration of pupil change; Indicates the average value of pupil change acceleration within a preset detection period; represents the standard deviation of pupil acceleration; Through the solution provided by this embodiment, the pupil change speed, pupil change direction and pupil change acceleration are taken into consideration, and the eyeball scaling index is first obtained by designing a formula calculation, so that the eyeball scaling index is used to reflect the eyeball movement. On the one hand, the evaluation of the eyeball movement can be made more comprehensive and accurate. On the other hand, the design formula can digitize complex content, so as to more simply reflect the eyeball movement.

[0097] In some embodiments, the last inspection data is determined based on the current detection mode; the last eyelid opening and closing degree, the last eyeball zoom degree and the last pupil change are determined based on the last inspection data; the eye change trend is determined based on the eyelid opening and closing degree, eyeball movement, pupil changes, the last eyelid opening and closing degree, the last eyeball zoom degree and the last pupil change; based on the eye change trend, the current eye disease state is determined.

[0098] The last examination data can be understood as the examination data obtained from the last eye test before the current examination, which may include the last eyelid opening and closing degree, eyeball movement, and pupil changes, that is, the last eyelid opening and closing degree, the last eyeball zoom degree, and the last pupil change.

[0099] Specifically, according to the current detection mode, the last examination data matching the current detection mode is extracted from the database, the last examination data is classified, and the last eyelid opening and closing degree, the last eyeball zoom degree and the last pupil change are determined. The eyelid opening and closing degree, eyeball movement and pupil change at the current moment are compared with the corresponding last eyelid opening and closing degree, the last eyeball zoom degree and the last pupil change, respectively, to determine the trend of eye changes compared with the last time under the current detection mode, that is, the eye change trend. For example, under the same detection mode, the eyelid opening and closing degree is larger than the last eyelid opening and closing degree, and based on the patient's condition, if the eye opening and closing degree gradually increases, it means that the condition has improved. At this time, it can be considered that the eye change trend is changing in a good direction. Based on this eye change trend, it is determined whether the current eye disease has improved compared with the last time, and then combined with the initial diagnosis by the doctor, the current eye disease status is determined.

[0100] The solution provided in this embodiment integrates data from multiple dimensions such as eyelid opening and closing, eyeball movement, pupil changes, etc., providing a solid foundation for a comprehensive assessment of eye conditions. By comparing the current test data with the last test data, the changing trend of the eye condition can be accurately captured. By analyzing the trend of eye changes, the current state of the eye disease can be determined, providing a basis for subsequent judgment of whether the eye disease has improved.

[0101] Figure 3 A schematic diagram of a system for monitoring the progression of thyroid-related eye disease provided in one embodiment of the present application is shown in FIG. Figure 3 As shown, the thyroid-related eye disease progression monitoring system 300 of this embodiment includes: a feature determination module 301 , a feature analysis module 302 , a pupil analysis module 303 , a state analysis module 304 , and a symptom analysis module 305 .

[0102] The feature determination module 301 is used to obtain the eye monitoring video taken by the camera device, analyze the eye monitoring video, and determine the eye features; A feature analysis module 302 is used to analyze the eye features and determine the degree of eyelid opening and closing and eye movement; The pupil analysis module 303 is used to determine pupil changes according to the eye movement; The state analysis module 304 determines the current eye disease state according to the eyelid opening and closing degree, the eyeball movement and the pupil change; The symptom analysis module 305 is used to obtain and analyze historical eye disease diagnosis and treatment records to obtain historical eye disease status; compare the current eye disease status with the historical eye disease status to determine whether the thyroid-related eye disease has improved.

[0103] Optionally, the feature determination module 301 is specifically configured to: Acquire and analyze a starting eye image, and determine whether there is eye deviation based on the image analysis result; If it is determined that there is eye deviation, the eye deviation angle is determined based on the image analysis result; The eye difference is determined according to the eye deviation angle and the historical eye disease diagnosis and treatment records, and the image capture method is adjusted according to the eye difference, and the eye monitoring video is obtained according to the adjusted image capture method.

[0104] Optionally, the eye detector includes a camera device; the feature determination module 301 is specifically used for: monitoring the displacement change of the eye detector in real time, and determining whether the spatial position of the eye detector changes according to the displacement change; If it is determined that there is a spatial position change, determining a horizontal position change amplitude according to the spatial position change; Determining whether to start the camera device according to the horizontal position change amplitude; If it is determined to start the camera device, acquiring and screening the captured image of the camera device to obtain the initial eye image; Analyze the initial eye image, and extract eye features according to the image analysis result; According to the eye features, a structure boundary is determined, and according to the structure boundary, the eye features in each structure are analyzed to determine whether there is an eye offset.

[0105] Optionally, the feature determination module 301 is specifically configured to: determining an eyelid region based on the structure boundaries and eye features within each structure; Analyzing eye features of the eyelid region to determine an eyelid shadow area; Determining the eyelid shadow area according to the eyelid shadow area; Obtaining current lighting conditions and the spatial position of the camera device; Determining a desired eyelid shadow area according to the current lighting condition and the spatial position; Comparing the eyelid shadow area with the expected eyelid shadow area to determine whether the eyelid shadow area is greater than the expected eyelid shadow area; If the eyelid shadow area is greater than the expected eyelid shadow area, it is determined that eye deviation exists.

[0106] Optionally, the feature determination module 301 is specifically configured to: Analyzing the eye monitoring video to determine changes in light and dark; Determining a current detection mode according to the light and dark changes; According to the current detection mode, the eye monitoring video is analyzed frame by frame to determine eye features.

[0107] Optionally, the feature analysis module is specifically used to: Analyzing the eye features to determine the eyelid contour and eyeball area; Determining the relative position between the eyelid contour and the eyeball area according to the feature vector of the eye feature; Determine the opening and closing ratio of the eyelid according to the relative position, and determine the opening and closing degree of the eyelid according to the opening and closing ratio; Determine the pupil position in each frame of video according to the eye features; According to the pupil position in each frame of video, the pupil change speed, pupil change direction and pupil change acceleration are determined; The eye movement condition is determined according to the pupil change speed, the pupil change direction and the pupil change acceleration.

[0108] Optionally, the pupil analysis module 303 is specifically used for: Based on the eye movement, real-time acquisition of environmental changes; determining a pupil inspection mode according to the environmental changes; Determining the real-time position of the pupil according to the eye movement; Determining a real-time pupil diameter based on the real-time pupil position and the eye feature; The pupil change is determined according to the pupil inspection mode and the real-time pupil diameter.

[0109] Optionally, the feature analysis module 302 is specifically used to: ; in, represents the eye movement index; represents the pupil change speed; A weight coefficient representing the pupil change speed; Indicates the average value of the pupil change speed within a preset detection period; represents the standard deviation of the pupil change speed; A weight coefficient representing the direction of pupil change; Indicates the quantified direction of pupil change; A weight coefficient representing the acceleration of pupil change; represents the pupil change acceleration; represents the acceleration mean of the pupil change acceleration within a preset detection period; represents the standard deviation of the pupil change acceleration; The eye movement condition is determined according to the eye movement index.

[0110] Optionally, when determining the current eye disease state according to the eyelid opening and closing degree, the eyeball movement, and the pupil change, the state analysis module 304 is specifically used to: Determining the last inspection data according to the current inspection mode; Determining the last eyelid opening and closing degree, the last eyeball zooming degree and the last pupil change according to the last examination data; Determining the eye change trend according to the eyelid opening and closing degree, the eyeball movement condition, the pupil change condition, the last eyelid opening and closing degree, the last eyeball zooming degree and the last pupil change condition; The current state of the eye disease is determined based on the eye change trend.

[0111] The system of this embodiment can be used to execute the method of any of the above embodiments. The implementation principles and technical effects are similar and will not be described in detail here.

Claims

1. A method for monitoring the progression of thyroid-related eye disease, It is characterized in that The progress monitoring method is applied to an eye detector, the eye detector includes a camera device and a data processing chip, and the method is applied to the data processing chip; the method includes: Acquire the eye monitoring video captured by the camera device, analyze the eye monitoring video, and determine eye features; Analyzing the eye characteristics to determine the degree of eyelid opening and closing and eye movement; Determining pupil changes according to the eye movement; Determining the current eye disease state according to the eyelid opening and closing degree, the eyeball movement, and the pupil change; Obtain and analyze historical eye disease diagnosis and treatment records to obtain historical eye disease status; compare the current eye disease status with the historical eye disease status to determine whether the thyroid-related eye disease has improved.

2. The method according to claim 1, It is characterized in that The step of obtaining the eye monitoring video captured by the camera device includes: Acquire and analyze a starting eye image, and determine whether there is eye deviation based on the image analysis result; If it is determined that there is eye deviation, the eye deviation angle is determined based on the image analysis result; The eye difference is determined according to the eye deviation angle and the historical eye disease diagnosis and treatment records, and the image capture method is adjusted according to the eye difference, and the eye monitoring video is obtained according to the adjusted image capture method.

3. The method according to claim 2, It is characterized in that The step of acquiring and analyzing the initial eye image and determining whether there is eye deviation according to the image analysis result includes: monitoring the displacement change of the eye detector in real time, and determining whether the spatial position of the eye detector changes according to the displacement change; If it is determined that there is a spatial position change, determining a horizontal position change amplitude according to the spatial position change; Determining whether to start the camera device according to the horizontal position change amplitude; If it is determined to start the camera device, acquiring and screening the captured image of the camera device to obtain the initial eye image; Analyze the initial eye image, and extract eye features according to the image analysis result; According to the eye features, a structure boundary is determined, and according to the structure boundary, the eye features in each structure are analyzed to determine whether there is an eye offset.

4. The method according to claim 3, It is characterized in that Analyzing the eye features in each structure according to the structure boundary to determine whether there is eye deviation includes: determining an eyelid region based on the structure boundaries and eye features within each structure; Analyzing eye features of the eyelid region to determine an eyelid shadow area; Determining the eyelid shadow area according to the eyelid shadow area; Obtaining current lighting conditions and the spatial position of the camera device; Determining a desired eyelid shadow area according to the current lighting condition and the spatial position; Comparing the eyelid shadow area with the expected eyelid shadow area to determine whether the eyelid shadow area is greater than the expected eyelid shadow area; If the eyelid shadow area is greater than the expected eyelid shadow area, it is determined that eye deviation exists.

5. The method according to claim 1, It is characterized in that The analyzing the eye monitoring video to determine eye features includes: Analyzing the eye monitoring video to determine changes in light and dark; Determining a current detection mode according to the light and dark changes; According to the current detection mode, the eye monitoring video is analyzed frame by frame to determine eye features.

6. The method according to claim 5, It is characterized in that The analyzing of the eye features to determine the degree of eyelid opening and closing and eyeball movement includes: Analyzing the eye features to determine the eyelid contour and eyeball area; Determining the relative position between the eyelid contour and the eyeball area according to the feature vector of the eye feature; Determine the opening and closing ratio of the eyelid according to the relative position, and determine the opening and closing degree of the eyelid according to the opening and closing ratio; Determine the pupil position in each frame of video according to the eye features; According to the pupil position in each frame of video, the pupil change speed, pupil change direction and pupil change acceleration are determined; The eye movement condition is determined according to the pupil change speed, the pupil change direction and the pupil change acceleration.

7. The method according to claim 6, It is characterized in that Determining pupil changes according to the eye movement includes: Based on the eye movement, real-time acquisition of environmental changes; determining a pupil inspection mode according to the environmental changes; Determining the real-time position of the pupil according to the eye movement; Determining a real-time pupil diameter based on the real-time pupil position and the eye feature; The pupil change is determined according to the pupil inspection mode and the real-time pupil diameter.

8. The method according to claim 6, It is characterized in that The eye movement is determined according to the pupil change speed, the pupil change direction and the pupil change acceleration, referring to the following formula: ; in, represents the eye movement index; represents the pupil change speed; A weight coefficient representing the pupil change speed; Indicates the average value of the pupil change speed within a preset detection period; represents the standard deviation of the pupil change speed; A weight coefficient representing the direction of pupil change; Indicates the quantified direction of pupil change; A weight coefficient representing the acceleration of pupil change; represents the pupil change acceleration; represents the average value of the acceleration of pupil change within a preset detection period; represents the standard deviation of the pupil change acceleration; The eye movement condition is determined according to the eye movement index.

9. The method according to claim 6, It is characterized in that Determining the current eye disease state according to the eyelid opening and closing degree, the eyeball movement, and the pupil change includes: Determining the last inspection data according to the current inspection mode; Determining the last eyelid opening and closing degree, the last eyeball zooming degree and the last pupil change according to the last examination data; Determining the eye change trend according to the eyelid opening and closing degree, the eyeball movement condition, the pupil change condition, the last eyelid opening and closing degree, the last eyeball zooming degree and the last pupil change condition; The current state of the eye disease is determined based on the eye change trend.

10. A system for monitoring the progression of thyroid-related eye disease, It is characterized in that include: A feature determination module, used to obtain an eye monitoring video taken by a camera device, analyze the eye monitoring video, and determine eye features; A feature analysis module, used to analyze the eye features and determine the degree of eyelid opening and closing and eyeball movement; A pupil analysis module, used to determine pupil changes according to the eye movement; A state analysis module, which determines the current state of eye disease according to the degree of eyelid opening and closing, the eyeball movement, and the pupil change; The symptom analysis module is used to obtain and analyze historical eye disease diagnosis and treatment records to obtain historical eye disease status; compare the current eye disease status with the historical eye disease status to determine whether the thyroid-related eye disease has improved.

Images