Senile cataract detection device

Abstract

A senile cataract detection device is wirelessly connected to a background server, and comprises an account registration module for a user to register an account; the user login module is used for a registered user to perform login operation; the function selection module is used for a user to select a to-be-detected item; the interaction module is used for guiding the user to complete the corresponding detection item according to the selection of the user through the function selection module, and completing the corresponding information collection work; and the data processing module is used for analyzing the information collected by the interaction module to judge whether the user has the cataract risk or not. According to the senile cataract detection device provided by the invention, three characteristic indexes of contrast sensitivity, eyeball movement performance and eye appearance capable of reflecting senile cataract are combined, a portable senile cataract diagnosis device is designed, and preliminary assessment of disease risk is realized.

Description

Senile cataract detection device Technical Field

[0001] The present invention particularly relates to a device for detecting senile cataracts. Background Art

[0002] Senile cataracts, also known as age-related cataracts, are one of the most common types of cataracts, most commonly occurring in people over 50 years old, with the incidence increasing significantly with age. Senile cataracts are degenerative changes in the lens caused by aging and are the result of a combination of factors.

[0003] To ensure eye health, regular eye examinations are crucial for older adults. Currently, the clinical diagnosis of cataracts relies primarily on specialized ophthalmic equipment and physicians, including intraocular pressure testing, slit lamp examinations, iris projection, and fundus imaging. However, these procedures are time-consuming, costly, and involve relatively limited medical resources, making it difficult to prevent or avoid cataract-induced blindness.

[0004] In recent years, the use of artificial intelligence (AI) to diagnose and prevent senile cataracts has become a new trend in medical development. However, existing diagnostic systems based on smart terminals often rely on medical imaging, such as fundus imaging, to diagnose the disease. While this alleviates the time-consuming and labor-intensive nature of examinations, it still requires high-precision equipment, making it inconvenient and difficult to scale up.

[0005] Summary of the Invention

[0006] The present invention provides a device for detecting senile cataracts to solve the above-mentioned technical problems, and specifically adopts the following technical solutions:

[0007] A senile cataract detection device, the senile cataract detection device is wirelessly connected to a background server, and the senile cataract detection device comprises:

[0008] Account registration module, used for users to register accounts;

[0009] User login module, used for registered users to log in;

[0010] Function selection module, used for allowing users to select items to be tested;

[0011] An interactive module, configured to guide the user to complete corresponding test items according to the user's selection through the function selection module, and complete corresponding information collection work;

[0012] The data processing module is used to analyze the information collected by the interaction module to determine whether the user has the risk of cataract.

[0013] Furthermore, the items that the user can select through the function selection module include: contrast sensitivity detection, eye movement assessment and eye appearance feature assessment;

[0014] The data processing module is used to comprehensively analyze the test results of the above three test items to determine whether the user has the risk of cataract.

[0015] Furthermore, the interaction module includes:

[0016] An image display unit, configured to display a corresponding test image according to a test item;

[0017] An image acquisition unit, used to acquire image information of the user when completing the corresponding project;

[0018] A distance detection unit, configured to detect the distance between the user and the mobile client;

[0019] The voice playback unit is used to issue voice prompts to guide users during testing.

[0020] Furthermore, when the user selects contrast sensitivity detection, the distance detection unit and the voice playback unit guide the user to move to the correct position, and then the image display unit sequentially displays a group of graphics with colors from dark to light moving along a specified motion trajectory, where the depth of the icons corresponds to different contrast sensitivity values. The image acquisition unit captures a first video image of the user when observing the moving graphics, and the data processing module determines the user's contrast sensitivity value based on the acquired first video image, and determines whether there is an abnormality based on whether the contrast sensitivity value is less than a preset sensitivity threshold.

[0021] Furthermore, the specific method by which the data processing module determines the contrast sensitivity value of the user based on the acquired first video image is:

[0022] Preprocessing the first video image;

[0023] Identifying a first facial region in the first video image;

[0024] Detecting first facial feature points from the first facial area;

[0025] Calculating a first gaze vector for each eye using a first eye feature point in the first facial feature points;

[0026] The coordinates of the line of sight mapped on the screen are calculated based on the first gaze vector, and then a scatter plot is drawn in chronological order to represent the line of sight position of each frame of the image. A straight line is fitted based on the scatter plot. The line and the movement trajectory of the corresponding image are judged to determine whether the two are consistent to determine whether the user can achieve the corresponding contrast sensitivity value, and finally the user's contrast sensitivity value is determined.

[0027] Furthermore, when the user selects eye movement assessment, the distance detection unit and the voice playback unit guide the user to move to the correct position, and then the voice playback unit issues a scanning instruction, and the image acquisition unit captures a second video image of the user when completing the relevant scanning instruction, and the data processing module performs eye movement assessment on the user based on the acquired second video image.

[0028] Furthermore, the specific method of the data processing module performing eye movement assessment on the user based on the acquired second video image is:

[0029] preprocessing the second video image;

[0030] identifying a second facial region in the second video image;

[0031] detecting second facial feature points from the second facial region;

[0032] The second eye feature point in the second facial feature point is used to calculate the second gaze vector of each eye at the start and end times of the stable gaze period, and the difference between the second gaze vectors at the start and end times is used to approximate the amplitude change of the binocular gaze. If the difference between the binocular gazes is less than a given threshold, the eye movement is considered normal. Conversely, if the difference in the amplitude of the binocular gazes is too large, the eye movement is considered abnormal.

[0033] Furthermore, when the user selects eye appearance feature evaluation, the distance detection unit and the voice playback unit guide the user to move to the correct position, and then the image acquisition unit acquires the user's eye image, and the data processing module determines whether the user has cataract appearance features based on the acquired eye image.

[0034] Furthermore, the specific method by which the data processing module determines whether the user has cataract appearance characteristics based on the acquired human eye image is:

[0035] The data processing module identifies the human eye image through the trained model, obtains a classification result, and determines whether the human eye image has cataract appearance features.

[0036] Furthermore, the interactive visual impairment intelligent detection system based on dynamic gaze behavior analysis further comprises:

[0037] User identification module, when the user logs in through the user login module, the image acquisition unit is further used to collect the user's facial image, the user identification module is used to identify the facial image, and the user login module approves the user's login operation after the user identification module has recognized the facial image;

[0038] The cycle setting module is used to dynamically set a detection cycle based on the detection results of this user;

[0039] The prompt module is used to send a prompt message to the reserved mobile phone number when the user fails to perform the detection within the detection period set by the period setting module.

[0040] The benefit of the present invention lies in the fact that the provided senile cataract detection device innovatively combines contrast sensitivity, eye movement performance and eye appearance, the three characteristic indicators that can reflect senile cataracts, to construct an evaluation algorithm based on deep learning, and to design a portable senile cataract diagnosis device to achieve a preliminary assessment of the risk of the disease. No eye tracker is required, and only the camera records the user's eye movement behavior under different visual stimuli, while collecting images of the user's eye appearance, and uses a deep learning algorithm to analyze whether there are abnormalities in the user's visual performance and eye appearance. The device can simply and quickly assess whether elderly users are at risk of cataracts by combining dynamic and static analysis of eye conditions, helping elderly users to check and monitor their visual function status at home on a daily basis, and guiding whether they need to seek medical attention for further examination, thereby improving medical efficiency and reducing medical time and costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only 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.

[0042] FIG1 is a schematic diagram of a device for detecting senile cataracts according to the present invention;

[0043] FIG2 is a schematic diagram of an image display unit during sensitivity detection according to the present invention;

[0044] FIG. 3 is another schematic diagram of the image display unit during sensitivity detection according to the present invention. DETAILED DESCRIPTION

[0045] The following describes in detail embodiments of the present application. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.

[0046] As shown in Figure 1, a device for detecting cataracts in accordance with the present application is provided, which is used for users to carry out self-diagnosis of cataract risk in a portable manner. The device is wirelessly connected to a backend server and exchanges data with the backend server.

[0047] In an embodiment of the present application, the senile cataract detection device includes: an account registration module, a user login module, a function selection module, an interaction module and a data processing module. Among them, the account registration module is used for users to register accounts. The user login module is used for registered users to perform login operations. The function selection module is used for users to select items to be tested. The interaction module is used to guide users to complete corresponding test items according to the user's selection through the function selection module, and complete the corresponding information collection work. The data processing module is used to analyze the information collected by the interaction module to determine whether the user has the risk of cataracts. Through the above-mentioned senile cataract detection device, elderly users do not need to go to the hospital, and can directly complete portable self-examination through the detection device to obtain test results.

[0048] In an embodiment of the present application, the senile cataract detection device is preferably a mobile smart terminal, such as a mobile phone, a smart tablet, etc. The interactive module includes an image display unit, an image acquisition unit, a distance detection unit, and a voice playback unit. The image display unit is used to display the corresponding test image according to the test item. The image acquisition unit is used to collect image information of the user when completing the corresponding item. The distance detection unit is used to detect the distance between the user and the mobile client. The voice playback unit is used to issue voice prompts to guide the user in the test.

[0049] The function selection module allows users to select from a range of test items, including contrast sensitivity testing, eye movement assessment, and ocular appearance assessment. The data processing module analyzes the results of these three tests to determine whether the user is at risk for cataracts.

[0050] In addition to impaired vision, cataracts can also lead to decreased contrast sensitivity and color perception. Therefore, contrast sensitivity can be used as a screening indicator for disease risk assessment. When a user selects contrast sensitivity testing, the distance detection unit and voice playback unit guide the user to the correct position. The image display unit then displays a series of graphics in descending order of color, as shown in Figures 2 and 3, moving along a specified motion trajectory—in this application, a linear motion. The depth of the icons corresponds to different contrast sensitivity values. The image acquisition unit captures a first video image of the user observing the moving graphics. The data processing module determines the user's contrast sensitivity value based on the captured first video image and determines whether an abnormality exists based on whether the contrast sensitivity value is less than a preset sensitivity threshold. Specifically, the determination is based on whether the user can accurately follow the movement of the image displayed on the image display unit. If the user is unable to follow the movement of an image of a certain depth, it indicates that the user cannot obtain the sensitivity value corresponding to that depth. Ultimately, the user's achievable sensitivity limit is determined. This value is then compared with a preset sensitivity threshold. If the value is greater than the threshold, no abnormality exists. If it is less than the threshold, it means that the user's contrast sensitivity is abnormal and there may be a risk of cataracts.

[0051] As a preferred embodiment, the specific method for the data processing module to determine the contrast sensitivity value of the user based on the acquired first video image is:

[0052] Preprocess the first video image. First, preprocess the first video image data. Each graphic's duration of continuous movement is 10 seconds. To eliminate external interference, only the middle 6 seconds of the video data, i.e., 2s–8s, are used for gaze behavior analysis. The video segments are sampled at 33ms intervals, i.e., 1s of video will produce 30 frames. Each frame is then processed.

[0053] Identify the first face region in the first video image. Call OpenCV's existing face detection model and build a face detection model based on a convolutional neural network (CNN). Retrieve facial information by analyzing Haar features and perform face region detection on each frame. Frames where no faces are detected or where multiple people are present are discarded. Images where only a single face is detected are retained in the center of the screen. A rectangular box is rendered at the face location to represent the face, completing face detection.

[0054] The first facial feature points are detected from the first facial region. The appropriate facial region of each frame is further processed as input. A facial landmark location algorithm based on a regression tree model is used to learn the features of the facial image. The model details are fitted to the coordinates of the facial structure to generate 68 facial landmarks to extract high-level facial information. The input facial image is then corrected based on the locations of the landmarks, such as the facial contour points, eye corners, pupils, and mouth corners, to eliminate errors caused by different head postures.

[0055] The first gaze vector of each eye is calculated using the first eye feature point in the first facial feature point. Specifically, by analyzing eye movements, information such as user intention, behavior, and attention can be obtained. In order to accurately capture eye movements, it is necessary to further refine the eye feature points based on the facial feature points that have been obtained. The constrained local neural fields (CLNF) algorithm is used to further analyze the key locations of the eyes, detect key points that reflect more details such as the eyelids, irises, and pupils, and then calculate the first gaze vector of each eye based on local features such as the pupil, eye corner, and eyeball position.

[0056] A ray is emitted from the origin in the camera coordinate system, passing through the pupil center in the image plane. The intersection with the eye area is calculated and recorded as the pupil position in 3D camera coordinates. Based on the relative positions of the 3D eye center and pupil, the gaze direction at that moment is estimated through vector calculation. This device uses an eye appearance-based gaze estimation algorithm, which estimates gaze based on the relative position of the eye shape and pupil. This method has low hardware requirements and can be implemented on platforms without high-resolution cameras or additional light sources.

[0057] The coordinates of the gaze mapped on the screen are calculated based on the first gaze vector. Then, a scatter plot is drawn in chronological order to represent the gaze position of each frame of the image. A straight line is fitted based on the scatter plot. The consistency between the straight line and the movement trajectory of the corresponding image is judged to determine whether the user can achieve the corresponding contrast sensitivity value, and finally the user's contrast sensitivity value is determined.

[0058] Contrast sensitivity is a sensitive indicator of eye disease and can, to a certain extent, predict visual impairments that visual acuity alone cannot. Based on the eye movement trajectory simulated in the previous steps, we can determine whether the user can distinguish moving icons on the screen.

[0059] All icons are pre-set to move in different directions. If the fitted eye trajectory approximates a straight line and closely matches the icon's movement trajectory, it can be assumed that the user can see the corresponding contrast sensitivity value. Conversely, if the user's eye trajectory changes erratically and a consistent straight line cannot be fitted, it is assumed that the corresponding contrast sensitivity value cannot be seen. A contrast sensitivity threshold is pre-set. If the user cannot see contrast sensitivity icons below this threshold, it is considered that there may be a risk of cataracts.

[0060] Eye movements are a natural reflex caused by visual changes induced by visual stimulation. They are an important component of visual perception and can more objectively reflect a person's visual ability. This application utilizes gaze estimation technology, which does not require any complex hardware such as an eye tracker. Instead, it uses a camera to capture the user's responses while performing visual tasks and analyzes the captured video to achieve gaze estimation.

[0061] When the user selects eye movement assessment, the distance detection unit and the voice playback unit guide the user to move to the correct position. Then the voice playback unit issues a scanning instruction, and the image acquisition unit captures a second video image of the user when completing the relevant scanning instruction. The data processing module performs eye movement assessment on the user based on the acquired second video image.

[0062] The specific method for the data processing module to evaluate the user's eye movement based on the acquired second video image is:

[0063] Preprocess the second video image. Identify a second facial region in the second video image. Detect second facial feature points from the second facial region. Calculate the second gaze vectors of each eye at the start and end times of a stable gaze period using the second eye feature points within the second facial feature points. Use the difference between the second gaze vectors at the start and end times to approximate the change in binocular gaze amplitude. If the difference between the binocular gaze amplitudes is less than a given threshold, the eye movement is considered normal. Conversely, if the difference in binocular gaze amplitude is too large, the eye movement is considered abnormal.

[0064] Unlike contrast sensitivity testing, which simulates gaze trajectories, eye movement assessment focuses on abnormal saccadic behavior. Following voice guidance from the device, the user performs large sweeps, such as maintaining a fixed head position while making large shifts of gaze from left to right or from top to bottom. By calculating the amplitude of each eye movement, the left and right eye movements are estimated and compared to analyze whether abnormal saccadic behavior exists.

[0065] When the user chooses eye appearance feature evaluation, the distance detection unit and voice playback unit guide the user to move to the correct position, and then the image acquisition unit captures the user's eye image. The data processing module determines whether the user has cataract appearance features based on the acquired eye image.

[0066] As a preferred embodiment, the specific method for the data processing module to determine whether the user has the appearance characteristics of cataracts based on the acquired human eye image is: the data processing module recognizes the human eye image through a trained model, obtains a classification result, and determines whether the human eye image has the appearance characteristics of cataracts.

[0067] This application aims at abnormal features of the lens of cataract patients, such as local white or brown turbidity, corneal turbidity, and pupil constriction under strong light. It starts with the eye appearance image and only analyzes the eye appearance image without the need for professional equipment to collect eye medical images. First, collect intuitive pictures of the eyes. Build a CNN model to detect the position of the eyes and crop them into images of a given size as input data. Based on a large amount of collected eye appearance image data, a ResNet model is trained to learn and extract intuitive features of the eyes. The feature vector is then input into the SVM classification model for binary classification. The user's eye image captured by the rear camera is analyzed to see if there are abnormal appearance features such as lens turbidity or corneal turbidity, and the risk assessment results for early cataracts are output.

[0068] As a preferred embodiment, an interactive intelligent visual impairment detection system based on dynamic gaze behavior analysis further includes: a user identification module and a cycle setting module. When a user logs in through the user login module, the image acquisition unit is further configured to capture the user's facial image. The user identification module is configured to identify the facial image. The user login module approves the user's login operation after the user identification module successfully identifies the user. The cycle setting module is configured to dynamically set a detection cycle based on the user's test results. The prompt module is configured to send a prompt message to a reserved mobile phone number if the user fails to undergo a test within the detection period set by the cycle setting module. Specifically, the cycle setting module dynamically sets an appropriate detection cycle based on the user's test results, and the set detection cycle is determined based on the user's test assessment results. Preferably, if the user's test results indicate that the user is not at risk of cataracts, the cycle setting module can set a longer detection cycle. If the user's test results indicate that the user is at risk or is close to being at risk, the cycle setting module can set a shorter detection cycle. This dynamic cycle generation method allows for targeted settings based on the user's specific circumstances.

[0069] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the above embodiments do not limit the present invention in any form, and any technical solutions obtained by equivalent replacement or equivalent transformation fall within the scope of protection of the present invention.

Claims

1. A senile cataract detection device, the senile cataract detection device is wirelessly connected to a background server, characterized in that: The senile cataract detection device comprises: Account registration module, used for users to register accounts; User login module, used for registered users to log in; Function selection module, used for users to select items to be tested; An interactive module, used to guide the user to complete the corresponding test items according to the selection made by the user through the function selection module, and complete the corresponding information collection work; The data processing module is used to analyze the information collected by the interaction module to determine whether the user has a risk of cataract.

2. The interactive visual impairment intelligent detection system based on dynamic gaze behavior analysis according to claim 1 is characterized in that: The items that the user can select through the function selection module include: contrast sensitivity detection, eye movement assessment and eye appearance feature assessment; The data processing module is used to comprehensively analyze the test results of the above three test items to determine whether the user has the risk of cataract.

3. The interactive visual impairment intelligent detection system based on dynamic gaze behavior analysis according to claim 2 is characterized in that: The interaction module comprises: An image display unit, used to display a corresponding test image according to a test item; An image acquisition unit, used to acquire image information of the user when completing the corresponding project; A distance detection unit, used to detect the distance between the user and the mobile client; The voice playback unit is used to issue voice prompts to guide users in testing.

4. The interactive visual impairment intelligent detection system based on dynamic gaze behavior analysis according to claim 3 is characterized in that: When the user selects contrast sensitivity detection, the distance detection unit and the voice playback unit guide the user to move to the correct position, and then the image display unit sequentially displays a group of graphics with colors from dark to light moving along a specified motion trajectory, wherein the depth of the icons corresponds to different contrast sensitivity values, and the image acquisition unit acquires a first video image of the user when observing the moving graphics, and the data processing module determines the contrast sensitivity value of the user based on the acquired first video image, and determines whether there is an abnormality based on whether the contrast sensitivity value is less than a preset sensitivity threshold.

5. The interactive visual impairment intelligent detection system based on dynamic gaze behavior analysis according to claim 4 is characterized in that: The specific method by which the data processing module determines the contrast sensitivity value of the user according to the acquired first video image is: Preprocessing the first video image; Identifying a first face region in the first video image; Detecting first facial feature points from the first facial area; Calculating a first gaze vector of each eye respectively through a first eye feature point among the first facial feature points; The line of sight coordinates mapped on the screen are calculated based on the first gaze vector, and then a scatter plot is drawn in chronological order to represent the line of sight position of each frame of the image. A straight line is fitted based on the scatter plot, and whether the straight line and the moving trajectory of the corresponding image are consistent is judged to determine whether the user can reach the corresponding contrast sensitivity value, and finally the user's contrast sensitivity value is determined.

6. The interactive visual impairment intelligent detection system based on dynamic gaze behavior analysis according to claim 3 is characterized in that: When the user selects eye movement assessment, the distance detection unit and the voice playback unit guide the user to move to the correct position, and then the voice playback unit issues a scanning instruction, and the image acquisition unit captures a second video image of the user when completing the relevant scanning instruction, and the data processing module performs eye movement assessment on the user based on the acquired second video image.

7. The interactive visual impairment intelligent detection system based on dynamic gaze behavior analysis according to claim 6, characterized in that: The specific method of the data processing module evaluating the user's eye movement according to the acquired second video image is: Preprocessing the second video image; Identifying a second face region in the second video image; Detecting second facial feature points from the second facial area; The second gaze vector of each eye at the start and end times of a time period in which the line of sight is stable is calculated using the second eye feature point in the second facial feature point, and the difference between the second gaze vectors at the start and end times is used to approximately represent the amplitude change of the line of sight of the two eyes. If the difference between the two eyes is less than a given threshold, the eye movement is considered normal. Conversely, if the difference in the amplitude of the two eyes' movement is too large, the eye movement is considered abnormal.

8. The interactive visual impairment intelligent detection system based on dynamic gaze behavior analysis according to claim 3 is characterized in that: When the user selects eye appearance feature evaluation, the distance detection unit and the voice playback unit guide the user to move to the correct position, and then the image acquisition unit acquires the user's eye image. The human eye image obtained is used to determine whether the user has cataract appearance features.

9. The interactive visual impairment intelligent detection system based on dynamic gaze behavior analysis according to claim 8, characterized in that: The specific method by which the data processing module determines whether the user has cataract appearance features according to the acquired human eye image is: The data processing module recognizes the human eye image through the trained model, obtains a classification result, and determines whether the human eye image has cataract appearance features.

10. The interactive visual impairment intelligent detection system based on dynamic gaze behavior analysis according to claim 3, characterized in that: The interactive visual impairment intelligent detection system based on dynamic gaze behavior analysis also includes: A user identification module, when a user logs in through the user login module, the image acquisition unit is further used to acquire a facial image of the user, the user identification module is used to identify the facial image, and the user login module approves the user's login operation after the user identification module has identified the facial image; The cycle setting module is used to dynamically set a detection cycle according to the detection result of this user; The prompt module is used to send a prompt message to the reserved mobile phone number when the user fails to perform the detection within the detection period set by the period setting module.

Images