Artificial intelligence eye disease screening service method and system

Abstract

The invention relates to the field of eye disease screening, in particular to an artificial intelligence eye disease screening service method and system, and the system comprises an imaging assembly which is used for obtaining an eye medical image of a patient; the image analysis module comprises a convolutional neural network, and the convolutional neural network is used for inputting eye images and outputting eye disease types; the convolutional neural network is also used for generating identification features for judging eye diseases and outputting relative positions of the identification features in the eye medical image; the imaging auxiliary module is used for performing key image extraction in the eye medical image acquired by the imaging assembly, performing monitoring based on a relative position of a preset identification feature in the medical image, performing gray scale gradient value judgment on the image at the relative position, and when an area meeting the gray scale gradient value is greater than a preset percentage, outputting the key image; and intercepting the image as a key image. According to the technical scheme, the medical image acquisition process is enhanced and accelerated through the imaging auxiliary module.

Description

Technical Field

[0001] The present invention relates to the field of ophthalmopathy screening, and specifically relates to an artificial intelligence ophthalmopathy screening service method and system. Background Art

[0002] Traditional methods for children's vision screening mainly rely on professional equipment in hospitals and ophthalmologists. Although the accuracy is high, the operation is complex and requires the high cooperation of children. In recent years, the rapid development of artificial intelligence technology has brought revolutionary changes to the field of medical health. The application of AI technology in the field of ophthalmology, such as the automatic detection of retinopathy, diabetic retinopathy, etc., has shown higher accuracy and efficiency than traditional methods. The AI screening model can identify children's ophthalmopathies by analyzing photos taken by smartphones, providing a convenient and early identification means for families, etc.

[0003] In the prior art, for example, the invention with the patent publication number CN109528155B discloses an intelligent screening system applicable to high myopia complicated with open-angle glaucoma and its establishment method, which uses models such as a glaucoma OCT deep learning model, a visual field deep learning model, and a fundus photo deep learning model to analyze and identify the medical images of patients' eyes; CN112084961B discloses an artificial intelligence ophthalmopathy screening and diagnosis and treatment system based on an ophthalmic robot, which completes the acquisition of eye images by setting an ophthalmic robot. However, the current number of ophthalmologists is far from enough. A team of four people for screening children's ophthalmopathies can screen about 600 people a day, while about 60,000 - 70,000 children need to be screened every year. Artificial intelligence can use the collected medical images for judgment, reducing the professional knowledge required for screening and reducing the manual workload, etc. The analysis process of artificial intelligence using medical images can be non-real-time, that is, it can collect separately and analyze the medical images at a time period other than the acquisition time, thereby greatly improving the work efficiency and increasing the number of screenings per day.

[0004] When performing children's screening, children have low attention concentration and unpredictable behaviors. At the same time, children are unable to fixate for a long time and have small pupils, which easily cause problems such as blurred images and image occlusion. Therefore, an artificial intelligence ophthalmopathy screening service method and system for processing and optimizing the acquisition efficiency of fundus photos, OCT, etc. during children's screening are needed. Summary of the Invention

[0005] To solve the above problems, the present invention provides an artificial intelligence ophthalmopathy screening service method and system for enhancing and accelerating the medical image acquisition process through an imaging assistance module.

[0006] To achieve the above object, the technical solution of the present invention is as follows: An artificial intelligence ophthalmopathy screening service system, comprising:

[0007] Imaging component: used to obtain medical images of a patient's eyes;

[0008] Image analysis module: includes a convolutional neural network, which is trained based on medical images of the eyes under various eye diseases. The convolutional neural network is used to input an eye image and output the type of eye disease; the convolutional neural network is also used to generate identification features for judging various eye diseases and output the relative positions of the identification features within the medical images of the eyes.

[0009] Imaging assistance module: used to extract key images during the acquisition of medical images of the eyes by the imaging component. The imaging assistance module independently monitors based on the relative positions of preset identification features within the medical images, and judges the gray gradient values of the images at the relative positions. When the area satisfying the gray gradient value is greater than the preset percentage, it is intercepted as a key image; after the key image is intercepted, it is input into the convolutional neural network of the image analysis module, and the convolutional neural network outputs the type of eye disease based on the key image.

[0010] The following beneficial effects are achieved by adopting the above solution:

[0011] 1. In this solution, artificial intelligence is used to assist in eye disease screening. A large number of medical images of the eyes under various eye diseases are used for training, so as to use artificial intelligence analysis to assist doctors in screening and judgment. In addition to learning the judgment of eye diseases, the image analysis module is also used to extract the judgment features of various eye diseases, that is, identification features, so as to facilitate the imaging assistance module to extract key images.

[0012] 2. In this solution, the imaging assistance module can speed up the process of collecting medical images of the eyes. For example, children have low attention concentration and unpredictable behaviors, and it is difficult to keep the various regions of their medical images of the eyes clearly stable, and it is easy to produce blurred parts. The imaging assistance module can perform highly targeted image acquisition according to the preset identification features. First, the imaging assistance module will monitor each area used to judge eye diseases, and use the gray gradient value judgment to judge the image clarity. When the image is blurred, its color gamut mixing degree is high and the gray gradient is not obvious, while when the image is clear, the lines are obvious and there are a large number of areas at the edges of the lines where the gray gradient values meet the preset values. When a single identification feature of the patient meets the gray gradient requirement and clarity, it is intercepted. Therefore, there is no need to wait for the entire medical image to be clear, and only the images when only the identification feature is clear need to be collected separately, adapting to the scenario where children have low attention concentration, frequent actions and unpredictable behaviors.

[0013] 3. In this solution, after the features required for the image analysis module to recognize are collected, the acquisition object can be replaced, so that the diagnosis and acquisition are separated. The patient can obtain the screening result after waiting for the analysis of the image analysis module to be completed, and the analysis process and the acquisition process are asynchronous. Therefore, the speed of acquisition and screening is increased, and more patients can be screened within the same time.

[0014] Further, the number of preset identification features in the imaging assistance module is at least 1. After the imaging assistance module extracts the key images of each identification feature, the acquisition of the ophthalmic medical image by the imaging component ends.

[0015] Beneficial effects: The imaging assistance module can extract multiple identification features respectively, intercept and record them when the identification features are clear enough, and end the acquisition of the ophthalmic medical image by the imaging component after each identification feature is collected, so as to switch the acquisition target to the next patient.

[0016] Further, the imaging component includes one or more of a fundus camera, an OCT instrument, and an eye axis measuring instrument.

[0017] Beneficial effects: According to different types of eye diseases, different parts of the eye are detected by different instruments. Therefore, the imaging component can be one or more of a fundus camera, an OCT instrument, and an eye axis measuring instrument.

[0018] Further, the image analysis module is used to divide the relative positions of the identification features of each eye disease under different medical images according to the different types of ophthalmic medical images collected by the imaging component;

[0019] After the imaging assistance module extracts the identification features of each eye disease within a single type of medical image, the acquisition of the ophthalmic medical image of this type of medical image ends.

[0020] Beneficial effects: The ophthalmic medical images of patients can be detected with the cooperation of multiple instruments, and the identification features required by each instrument may be different. Therefore, the completion degree of the extraction of the identification features of each type of medical image is judged separately.

[0021] Further, the relative positions are the positions with reference to the left and right eye contours at a preset angle respectively.

[0022] Beneficial effects: The collected images may be rotated by a certain angle or be one of the left and right eyes. Therefore, the relative positions need to be the relative positions under the recognition of the eye contours rotated by a certain angle and distinguish between the left and right eyes.

[0023] Further, the imaging assistance module is used to pair and identify the outer contour of the eye and the position of the optic disc based on the Hough algorithm, judge the left and right eyes according to the orientation of the optic disc in the outer contour of the eye, and monitor the relative positions of the identification features after judging the left and right eyes.

[0024] Beneficial effects: The Hough algorithm is an algorithm for identifying and extracting images of specified shapes, so it has good performance in identifying the outer contour of the eye and the optic disc. The rotation angle of the eye contour can be restored based on the identified outer contour of the eye. At the same time, the positions of the optic discs in the left and right eyes are different, and the optic disc images are relatively obvious and easy to identify. Therefore, the positions of the left and right eyes can be determined by identifying the positions of the optic discs.

[0025] Furthermore, the imaging assistance module is also used to input the types of eye diseases. The imaging assistance module searches for the identifying features and relative positions of the eye diseases according to the input types of eye diseases, and completes the process of presetting the relative positions of the identifying features based on the search results.

[0026] Beneficial effects: Users can input the types of eye diseases according to the purpose of screening. The imaging assistance module will automatically search for and preset the corresponding identifying features and relative positions according to the input types of eye diseases.

[0027] Furthermore, it also includes a key image fusion module: which is used to fuse several intercepted key images, extract the identifying features in the key images, and generate a fused image containing the identifying features of the key images;

[0028] It also includes a remote consultation module: The remote consultation module is used to provide a communication interface for remotely connecting primary medical institutions and ophthalmology experts for remote consultation;

[0029] The fused image is used to input into the image analysis module for eye disease type analysis and for the remote consultation module to conduct remote consultation.

[0030] Beneficial effects: Due to limitations in the training model, the image analysis module may not be able to handle all types of diseases. Therefore, when the image analysis module fails to identify the type of eye disease, manual intervention is still required for diagnosis. The remote consultation module can build a remote communication foundation to support remote consultation between primary medical institutions and ophthalmology experts. In addition, since the image analysis module uses key images for analysis and the identifying features are scattered, a key image fusion module is set up to fuse each key image for easy reference by primary medical institutions and ophthalmology experts.

[0031] Furthermore, it also includes a trend prediction module: which is used to bind patient information to the patient's various ophthalmic medical images, the output information of the image analysis module, and the remote consultation information to generate a patient symptom change database. The convolutional neural network is trained based on the patient symptom change database, and the convolutional neural network is used to input patient information and ophthalmic medical images and output the patient symptom change trend.

[0032] Beneficial effects: The manifestations of different eye disease symptoms can map the development trends of future eye diseases. Therefore, during training, the information of the same patients in each screening is registered, and the results of each screening are bound. Then, based on the ophthalmic medical images of each patient, the corresponding long-term eye disease development process is learned, so that when identifying the type of eye disease, the development trend of the eye disease can be predicted.

[0033] An artificial intelligence-based eye disease screening service method includes the following steps:

[0034] Step 1: Set the types of eye diseases to be screened as required. The number of types of eye diseases is at least 1, and input the types of eye diseases into the imaging assistance module.

[0035] Step 2: Guide the target patient group to the imaging component to collect ophthalmic medical images.

[0036] Step 3: The imaging assistance module intervenes in the process of collecting medical images. According to the gray gradient value, it judges the identifiability of the landmark features. When a single landmark feature meets the identifiability, it is intercepted as a key image. When the interception of each landmark feature is completed, the collection of medical images of a single patient ends, and the patient is notified to wait for the screening result, and the target of collecting medical images is switched to the next patient.

[0037] Step 4: The obtained key images can be processed by the image fusion module or directly sent to the image analysis module. The image analysis module identifies and analyzes the eye diseases of the patients based on the received images, outputs the types of eye diseases and the trend of symptom changes, and notifies the patients of the screening results after the analysis is completed.

[0038] Step 5: When the image analysis module cannot identify due to the blurry received images, the patient is notified to collect medical images again. When the image analysis module cannot identify based on the landmark features, a remote consultation is established, and the fused images are sent to the primary medical institutions and ophthalmic experts participating in the consultation, and the patients are notified of the screening results according to the results of the remote consultation.

[0039] Beneficial effects: During the screening, the image analysis module is used for screening assistance, thereby reducing the manual workload, and splitting the collection of medical images and the analysis of symptoms, so that the screening work of a single patient only needs to complete the collection of images, without waiting for the process of judging the eye symptoms, improving the screening efficiency. At the same time, the imaging assistance module is used to intercept the landmark features required in the process of imaging ophthalmic medical images, so as to shorten the collection time and further improve the screening efficiency.

[0040] The additional aspects and advantages of the present invention will be partially given in the following description, partially become obvious from the following description, or be understood through the practice of the present invention. Description of the Drawings

[0041] Figure 1Schematic diagram of the modules of the artificial intelligence ophthalmopathy screening service system according to an embodiment of the present invention;

[0042] Figure 2 Schematic logical diagram of the artificial intelligence ophthalmopathy screening service system according to an embodiment of the present invention;

[0043] Figure 3 Schematic diagram of the steps of the artificial intelligence ophthalmopathy screening method according to an embodiment of the present invention. Detailed implementation manners

[0044] The technical solution of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0045] The following will be further described in detail through specific implementation manners:

[0046] As shown in the Figures 1 - 3 accompanying drawings: An artificial intelligence ophthalmopathy screening service system includes:

[0047] An imaging component: used to obtain medical images of the patient's eyes, and the imaging component includes one or more of a fundus camera, an OCT instrument, and an eye axis measuring instrument;

[0048] An image analysis module: includes a convolutional neural network, which is trained based on medical images of the eyes under various ophthalmopathies. The convolutional neural network is used to input an eye image and output the type of ophthalmopathy; the convolutional neural network is also used to generate identification features for judging various ophthalmopathies and output the relative positions of the identification features within the medical images of the eyes. The image analysis module is used to divide the relative positions of the identification features of various ophthalmopathies under different medical images according to the different types of medical images of the patient's eyes collected by the imaging component.

[0049] The relative positions are the positions with reference to the left and right eye contours at a preset angle respectively.

[0050] An imaging assistance module: used to extract key images during the acquisition of medical images of the eyes by the imaging component. The imaging assistance module listens respectively based on the relative positions of the preset identification features within the medical images. The number of the preset identification features in the imaging assistance module is at least 1. The images at the relative positions are judged for gray gradient values. When the area satisfying the gray gradient value is greater than a preset percentage, they are intercepted as key images. After the key images are intercepted, they are input into the convolutional neural network of the image analysis module, and the convolutional neural network outputs the type of ophthalmopathy based on the key images. After the extraction of the identification features of various ophthalmopathies within a single type of medical image is completed, the acquisition of the medical images of the eyes of this type of medical image is ended.

[0051] The imaging assistance module is used to pair and identify the outer contour of the eyes and the position of the optic disc based on the Hough algorithm, determine the left and right eyes according to the orientation of the optic disc relative to the outer contour of the eyes, and monitor the relative position of the identifying features after determining the left and right eyes.

[0052] Artificial intelligence is used to assist in the screening of eye diseases. A large number of ophthalmic medical images under various eye diseases are used for training, so as to use artificial intelligence analysis to assist doctors in screening and judgment. In addition to learning the judgment of eye diseases, the image analysis module is also used to extract the judgment features of various eye diseases, that is, the identifying features, so as to facilitate the imaging assistance module to extract key images.

[0053] The imaging assistance module can speed up the process of collecting ophthalmic medical images. For example, children have low attention concentration and unpredictable behaviors, and it is difficult to keep the various regions of their ophthalmic medical images clearly stable, and blurred parts are likely to occur. The imaging assistance module can perform image collection with strong purpose according to the preset identifying features. First, the imaging assistance module will monitor each region used to judge eye diseases, and use the gray gradient value to judge the image clarity. When the image is blurred, its color gamut mixing degree is high and the gray gradient is not obvious, while when the image is clear, the lines are obvious, and there are a large number of regions with gray gradient values meeting the preset values at the edges of the lines. When a single identifying feature of the patient meets the gray gradient requirement and clarity, it is intercepted. Therefore, there is no need to wait for the entire medical image to be clear, and only the images when only this identifying feature is clear need to be collected separately, which is suitable for the scenario where children have low attention concentration, frequent and unpredictable actions.

[0054] When judging the position of the identifying features, theoretically the eye structures are the same, so the identifying features will be in fixed positions. However, it is difficult for child patients to keep the ophthalmic medical images that can provide a fixed angle. Children may rotate the ophthalmic medical images by 0 - 12° for various reasons, making it difficult to judge the identifying features through the preset positions. Moreover, the left and right eyes are symmetric structures, and the positions of the internal identifying features are opposite, so it is also necessary to identify the differences between the left and right eyes in the ophthalmic medical images.

[0055] The Hough algorithm is a feature extraction technology widely used in image processing and computer vision, mainly used to detect shapes such as straight lines, circles, and ellipses in images. It maps the points in the image space to the parameter space and uses a voting mechanism to identify objects with specific shapes. Therefore, the Hough algorithm can be used to judge the outer contour of the ophthalmic medical image and the position of the optic disc. By identifying the outer contour of the eyes, it is convenient to restore the rotation angle of the eye contour. At the same time, the positions of the optic discs in the left and right eyes are different, and the optic disc images are relatively obvious and easy to identify. Therefore, the positions of the left and right eyes can be judged by identifying the position of the optic disc.

[0056] After the features required by the image analysis module are collected, the collection object can be changed, so that diagnosis and collection are separated. The patient can get the screening results while waiting for the image analysis module to complete the analysis. The analysis process and the collection process are carried out asynchronously. This speeds up the collection and screening process and completes the screening of more patients in the same time.

[0057] The imaging auxiliary module is also used to input the type of eye disease. The imaging auxiliary module searches for the identification features and relative positions of the eye disease according to the input type of eye disease, and completes the relative position preset process of the identification features according to the search results.

[0058] The user can input the type of eye disease according to the purpose of screening, and the imaging auxiliary module will automatically search and preset the corresponding identification features and relative positions based on the input type of eye disease.

[0059] Key image blending module: used to blend several captured key images, extract the identifying features in the key images, and generate a blended image containing the identifying features of the key images.

[0060] The remote consultation module is used to provide a communication interface for remote consultation between primary medical institutions and ophthalmologists.

[0061] The blended image is used to input into the image analysis module for eye disease type analysis and for the remote consultation module for remote consultation.

[0062] The image analysis module may not be able to handle all types of symptoms due to the limitations of the training model. Therefore, when the image analysis module cannot complete the identification of eye diseases, manual intervention is still required for diagnosis. The remote consultation module can build a remote communication foundation to support remote consultation between primary medical institutions and ophthalmologists. In addition, since the image analysis module uses key images for analysis and the identification features are scattered, a key image blending module is set up to blend the key images for primary medical institutions and ophthalmologists to review.

[0063] Trend prediction module: used to bind patient information to the patient's various ocular medical images, image analysis module output information, and remote consultation information to generate a patient symptom change database. The convolutional neural network is trained based on the patient symptom change database. The convolutional neural network is used to input patient information and ocular medical images to output the patient's symptom change trend.

[0064] The manifestation of different eye disease symptoms can reflect the development trend of eye diseases in the future. Therefore, in training, the same patient information in each screening is registered and bound to each screening result. Based on the eye medical images of each patient, the corresponding long-term eye disease development process is learned, so that when the eye disease type is identified, the development trend of the eye disease can be predicted.

[0065] An artificial intelligence ophthalmopathy screening service method, comprising the following steps:

[0066] Step 1, set the type of ophthalmopathy to be screened as required. The type of ophthalmopathy is at least 1, and input the type of ophthalmopathy into the imaging assistance module;

[0067] Step 2, guide the target patient group to the imaging component to collect ophthalmic medical images;

[0068] Step 3, the imaging assistance module intervenes in the process of collecting medical images, and judges the identifiability of the landmark features according to the gray gradient value. When a single landmark feature meets the identifiability, it is intercepted as a key image; when the interception of each landmark feature is completed, the collection of medical images of a single patient ends, notify the patient to wait for the screening result, and switch the target of medical image collection to the next patient;

[0069] Step 4, the obtained key images can be processed by the image fusion module or directly sent to the image analysis module. The image analysis module identifies and analyzes the ophthalmopathy of the patient based on the received images, outputs the type of ophthalmopathy and the trend of symptom changes, and notifies the patient of the screening result after the analysis is completed;

[0070] Step 5, when the image analysis module cannot identify due to the blurred received images, notify the patient to collect medical images again; when the image analysis module cannot identify based on the landmark features, establish a remote consultation, and send the fused images to the primary medical institutions and ophthalmology experts participating in the consultation, and notify the patient of the screening result according to the remote consultation result.

[0071] During the screening, the image analysis module is used for screening assistance, so as to reduce the manual workload, and split the collection of medical images and the analysis of symptoms, so that the screening work of a single patient only needs to complete the collection of images, without waiting for the process of judging the eye symptoms, improving the screening efficiency. At the same time, the imaging assistance module is used to intercept the landmark features required in the process of ophthalmic medical image imaging, so as to shorten the collection time and further improve the screening efficiency.

[0072] Obviously, the above embodiments are only examples clearly described and not limitations on the implementation manners. For those of ordinary skill in the art, other different forms of changes or variations can be made based on the above description. It is not necessary and impossible to list all the implementation manners here. And the obvious changes or variations derived therefrom are still within the protection scope of the present invention.

Claims

1. An artificial intelligence ophthalmopathy screening service system, characterized in that, Including: Imaging component: used to acquire medical images of the patient's eyes; Image analysis module: including a convolutional neural network, which is trained based on medical images of the eyes under various eye diseases. The convolutional neural network is used to input an eye image and output the type of eye disease; the convolutional neural network is also used to generate identifying features for judging various eye diseases and output the relative positions of the identifying features within the medical images of the eyes; Imaging assistance module: used to extract key images from the medical images of the eyes acquired by the imaging component. The imaging assistance module independently monitors based on the relative positions of the preset identifying features within the medical images and judges the gray gradient values of the images at the relative positions. When the area satisfying the gray gradient value is greater than the preset percentage, it is intercepted as a key image; After the key image is intercepted, it is input into the convolutional neural network of the image analysis module, and the convolutional neural network outputs the type of eye disease based on the key image.

2. The artificial intelligence ophthalmopathy screening service system according to claim 1, wherein The number of preset identifying features in the imaging assistance module is at least 1. After the imaging assistance module extracts the key images of each identifying feature, the acquisition of the medical images of the eyes by the imaging component ends.

3. The artificial intelligence ophthalmopathy screening service system according to claim 2, characterized in that The imaging component includes one or more of a fundus camera, an OCT instrument, and an eye axis measuring instrument.

4. The artificial intelligence ophthalmopathy screening service system according to claim 3, wherein The image analysis module is used to divide the relative positions of the identifying features of various eye diseases under different medical images according to the different types of medical images of the patient's eyes collected by the imaging component; After the imaging assistance module extracts the identifying features of each eye disease within a single type of medical image, the acquisition of the medical images of the eyes of this type of medical image ends.

5. The artificial intelligence ophthalmopathy screening service method and system according to claim 4, characterized in that The relative position is the position referenced by the left and right eye contours at a preset angle respectively.

6. The artificial intelligence ophthalmopathy screening service system according to claim 5, characterized in that, The imaging assistance module is used to pair and identify the outer contour of the eye and the position of the optic disc based on the Hough algorithm, judge the left and right eyes according to the orientation of the optic disc within the outer contour of the eye, and monitor the relative positions of the identifying features after judging the left and right eyes.

7. The artificial intelligence ophthalmopathy screening service system according to claim 6, wherein The imaging assistance module is also used to input the type of eye disease. The imaging assistance module searches for the identifying features and relative positions of the eye disease according to the input type of eye disease and completes the preset process of the relative positions of the identifying features according to the search results.

8. The artificial intelligence ophthalmopathy screening service system according to claim 7, characterized in that, It also includes a key image fusion module: used to fuse several intercepted key images, extract the identifying features within the key images, and generate a fused image containing the identifying features of the key images; It also includes a remote consultation module: the remote consultation module is used to provide a communication interface for remote docking between primary medical institutions and ophthalmology experts for remote consultation; The fused image is used to input into the image analysis module for eye disease type analysis and for the remote consultation module to conduct remote consultation.

9. The artificial intelligence ophthalmopathy screening service system according to claim 8, characterized in that, It also includes a trend prediction module: used to bind patient information to the patient's various medical images of the eyes, the output information of the image analysis module, and the remote consultation information to generate a patient symptom change database. The convolutional neural network is trained based on the patient symptom change database. The convolutional neural network is used to input patient information and medical images of the eyes and output the patient symptom change trend.

10. An artificial intelligence-based eye disease screening service method, which is based on the method of the artificial intelligence-based eye disease screening service system of claim 9, characterized in that, Including the following steps: Step 1, set the types of eye diseases to be screened according to needs. The types of eye diseases are at least 1, and input the types of eye diseases into the imaging assistance module; Step 2: Guide the target patient group to the imaging component for ophthalmic medical image acquisition; Step 3: The imaging assistance module intervenes in the medical image acquisition process, judges the recognizability of the identification features according to the gray gradient value. When a single identification feature meets the recognizability, it is intercepted as a key image; when all the identification features are intercepted, the medical image acquisition of a single patient is ended, the patient is notified to wait for the screening result, and the medical image acquisition target is switched to the next patient; Step 4: The acquired key images can be processed by the image fusion module or directly sent to the image analysis module. The image analysis module recognizes and analyzes the eye diseases of the patient based on the received images, outputs the types of eye diseases and the trend of symptom changes, and notifies the patient of the screening result after the analysis; Step 5: When the image analysis module cannot recognize due to the blurred received images, the patient is notified to perform medical image acquisition again; when the image analysis module cannot recognize based on the identification features, a remote consultation is established, and the fused images are sent to the primary medical institutions and ophthalmology experts participating in the consultation, and the patient is notified of the screening result according to the remote consultation result.

Images