Tear film lipid layer thickness detection method and system

Abstract

The invention relates to the technical field of medical detection, in particular to a tear film lipid layer thickness detection method and system, and the method comprises the following steps: collecting the interference image data of an ocular surface tear film through a special device, preprocessing the collected interference image data, carrying out the feature extraction of the preprocessed interference image, and obtaining the thickness of a tear film lipid layer. Extracting color, texture and shape features in the image, generating a feature vector, constructing and training a tear film lipid layer thickness prediction model by adopting a machine learning algorithm based on the feature vector, inputting interference image data of a to-be-detected tear film into the trained prediction model, calculating the thickness of the tear film lipid layer, and verifying a calculation result. The verified model and method are applied to clinical tear film lipid layer thickness detection to assist doctors in diagnosis and treatment of meibomian gland dysfunction. The tear film lipid layer thickness can be rapidly and accurately detected, manual operation errors are reduced, clinical diagnosis efficiency and accuracy are improved, and the method has remarkable clinical application value.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of medical detection, in particular to a tear film lipid layer thickness detection method and system. BACKGROUND

[0002] Tear film is a thin film structure covering the surface of the eyeball, which has the functions of lubricating the ocular surface, providing a smooth optical interface and resisting the invasion of external microorganisms. The tear film is composed of three layers: the innermost mucus layer, the middle water-like layer and the outermost lipid layer. The lipid layer is mainly composed of lipids secreted by the meibomian glands, which has the function of reducing tear evaporation and stabilizing the tear film. The thickness of the tear film lipid layer is an important indicator of the health status of the tear film, which directly affects the stability and function of the tear film. Therefore, accurately measuring the thickness of the tear film lipid layer is of great significance for the diagnosis and treatment of dry eye, meibomian gland dysfunction and other ocular surface diseases.

[0003] In the prior art, the detection of tear film lipid layer thickness mainly relies on interference imaging technology, which collects the interference image of the tear film through a white light interferometer, and analyzes the interference image using image processing technology to calculate the thickness of the lipid layer. However, the existing method has the following technical problems: first, the device is complex and expensive, which is not conducive to widespread application in clinical practice; second, the image processing algorithm is not perfect and is easily disturbed by noise, resulting in insufficient accuracy of the detection results; third, there is a lack of systematic verification means, making it difficult to ensure the reliability of the detection results. Traditional methods rely on manual operation, which is inefficient and prone to human error. SUMMARY

[0004] Based on the above purpose, the present application provides a tear film lipid layer thickness detection method and system.

[0005] A tear film lipid layer thickness detection method, comprising the following steps: S1, data acquisition: collecting the interference image data of the ocular surface tear film through a special device, the special device comprising a white light source, an interferometer, a camera and a computer system; S2, image preprocessing: preprocessing the collected interference image data, including image denoising, contrast enhancement and standardization processing; S3, feature extraction: extracting features from the preprocessed interference image, extracting color, texture and shape features in the image, generating a feature vector, the feature extraction including color histogram analysis, texture analysis and shape analysis, respectively extracting color information, texture pattern and shape features in the interference image; S4, model training: based on the extracted feature vector, a machine learning algorithm is used to construct and train a tear film lipid layer thickness prediction model; S5, thickness calculation: input the interference image data of the tear film to be detected into the trained prediction model, calculate the thickness of the tear film lipid layer, output the thickness value, and generate a detection report; S6, result verification: verify the calculation result by comparing the standard sample data and expert evaluation to verify the accuracy of the detection result; S7, clinical application: apply the verified model and method to clinical tear film lipid layer thickness detection to assist doctors in diagnosing and treating meibomian gland dysfunction.

[0006] Optionally, S1 specifically includes: White light source: used for illuminating the ocular surface tear film, providing uniform and stable light; Interferometer: used for generating interference images of the tear film, reflecting white light through the upper and lower interfaces of the tear film lipid layer to produce interference fringes; Camera: used for capturing interference images; Computer system: used for storing and processing collected interference image data.

[0007] Optionally, S2 specifically includes: S21, image denoising: using Gaussian filtering to denoise the collected interference image data; S22, contrast enhancement: using histogram equalization to enhance the contrast of the collected interference image data; S23, standardization processing: using image standardization method to standardize the collected interference image data.

[0008] Optionally, S3 specifically includes: S31, color histogram analysis: color histogram analysis describes the color features of the image by counting the pixel value distribution of each color channel in the image; S32, texture analysis: texture analysis extracts texture features by describing the spatial distribution relationship between pixels in the image; S33, shape analysis: shape analysis extracts shape features by describing the geometric shape features of objects in the image; S34, feature vector generation: combine the extracted color, texture and shape features into a feature vector.

[0009] Optionally, S4 specifically includes: S41, data preparation: combine the extracted color, texture and shape features into a feature vector, and label the corresponding tear film lipid layer thickness value to form a data set; S42, select machine learning algorithm: use convolutional neural network to build a tear film lipid layer thickness prediction model; S43, model training: training the convolutional neural network model using the dataset.

[0010] Optionally, the S5 specifically includes: S51, preprocessing: preprocessing the interferometric image data of the tear film to be detected to ensure that the image data quality meets the requirements of the prediction model, the preprocessing step including denoising, contrast enhancement and standardization processing; S52, feature extraction: extracting color, texture and shape features from the preprocessed interferometric image to generate a feature vector, the features including color histogram analysis, gray level co-occurrence matrix texture analysis and Fourier descriptor shape analysis; S53, calculating the thickness of the lipid layer of the tear film: inputting the generated feature vector into the trained tear film lipid layer thickness prediction model, and the model outputting the predicted thickness of the tear film lipid layer according to the input feature vector; S54, generating a detection report: generating a detection report according to the predicted thickness value, the report content including the thickness of the tear film lipid layer, the detection time and the related diagnostic suggestions.

[0011] Optionally, the S6 specifically includes: S61, standard sample data: preparing standard sample data, the standard sample data being sample data with pre-known thickness of the tear film lipid layer, usually measured by experts through other accurate methods (such as clinical evaluation or high-precision experimental equipment); S62, prediction model output result: calculating the thickness of the tear film lipid layer of the sample to be detected using the trained tear film lipid layer thickness prediction model, and recording the predicted value of each sample; S63, result comparison and analysis: comparing and analyzing the prediction results of the model with the standard sample data, calculating the error and statistical indicators between the prediction results and the standard values, and evaluating the accuracy of the model; S64, expert evaluation: inviting ophthalmologists to evaluate the prediction results of the samples; S65, model adjustment and optimization: adjusting and optimizing the tear film lipid layer thickness prediction model according to the feedback of error analysis and expert evaluation, to improve the accuracy and robustness of the model, the adjustment method including retraining the model, increasing the training data, optimizing the feature extraction method and adjusting the model parameters.

[0012] Optionally, the S7 specifically includes: S71, deployment: integrating the verified model and detection method into the ophthalmic diagnosis and treatment equipment of hospitals or clinics, ensuring that the system can run stably and be compatible with other diagnosis and treatment equipment, training medical personnel on operation, and ensuring that doctors and technical personnel are proficient in the use of the equipment; S72, application: when the patient is examined, the interference image of the tear film is collected using a special device, the collected interference image is preprocessed, the color, texture and shape features are extracted from the preprocessed interference image, the feature vector is generated, the feature vector is input into the trained tear film lipid layer thickness prediction model, and the thickness value of the tear film lipid layer is output.

[0013] Optionally, the S7 further comprises: S73, report generation: generate a patient detection report, including tear film lipid layer thickness, detection time and related diagnosis suggestions; S74, diagnosis and treatment: the doctor comprehensively evaluates the patient's meibomian gland function according to the detection report and clinical information.

[0014] A tear film lipid layer thickness detection system for implementing the above-mentioned tear film lipid layer thickness detection method, comprising the following modules: Data acquisition module: used for collecting interference image data of ocular surface tear film through a special device, the special device comprising a white light source, an interferometer, a high-definition camera and a computer system; Image preprocessing module: used for preprocessing the collected interference image data, including image denoising, contrast enhancement and standardization processing; Feature extraction module: used for feature extraction of the preprocessed interference image, extracting color, texture and shape features in the image, generating a feature vector, the feature extraction including color histogram analysis, texture analysis and shape analysis, respectively extracting color information, texture pattern and shape features in the interference image; Model training module: used for constructing and training a tear film lipid layer thickness prediction model based on the extracted feature vector using a machine learning algorithm; Thickness calculation module: used for inputting the interference image of the tear film to be detected into the trained prediction model, calculating the thickness of the tear film lipid layer, outputting the thickness value, and generating a detection report; Result verification module: used for verifying the calculation result by comparing standard sample data and expert evaluation to verify the accuracy of the detection result; Clinical application module: used for applying the verified model and method to clinical tear film lipid layer thickness detection to assist doctors in diagnosing and treating meibomian gland dysfunction.

[0015] The beneficial effects of the present application are: This invention improves the quality of interferometric image data by integrating multiple image preprocessing and feature extraction techniques. The comprehensive application of various feature extraction methods (color histogram analysis, gray-level co-occurrence matrix texture analysis, and Fourier descriptor shape analysis) ensures comprehensive capture and accurate description of image information. The use of convolutional neural networks to build a model for training and prediction effectively improves the prediction accuracy and robustness of tear film lipid layer thickness. Comparison with standard sample data and expert evaluation further ensures the accuracy and reliability of the detection results.

[0016] This invention modularizes and automates the tear film lipid layer thickness detection process. From data acquisition, image preprocessing, and feature extraction to thickness calculation, result verification, and report generation, each step is automated, reducing human error. Doctors and technicians can become proficient after brief training, improving the convenience and efficiency of clinical operations. Simultaneously, the system can quickly generate detailed reports containing tear film lipid layer thickness, detection time, and diagnostic recommendations, enabling doctors to quickly understand the patient's tear film health status and make timely diagnostic and treatment decisions.

[0017] This invention not only provides tear film lipid layer thickness detection but also generates personalized diagnostic suggestions and treatment plans based on the detection results. The integrated clinical application module assists doctors in diagnosing meibomian gland dysfunction, providing specific treatment suggestions such as warm compresses, meibomian gland massage, and drug therapy, thus improving the early diagnosis and intervention effectiveness of meibomian gland dysfunction and significantly enhancing its clinical application value. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only for this invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a schematic diagram of the method flow according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the system flow according to an embodiment of the present invention. Detailed Implementation

[0020] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. It should also be noted that, to make the embodiments more comprehensive, the following embodiments are the best and preferred embodiments, and those skilled in the art can use other alternative methods to implement some well-known technologies; moreover, the accompanying drawings are only for more specific description of the embodiments and are not intended to specifically limit the present invention.

[0021] It should be noted that the use of terms such as "an embodiment," "an embodiment," "an exemplary embodiment," and "some embodiments" in the specification indicates that the described embodiment may include a specific feature, structure, or characteristic, but not every embodiment necessarily includes that specific feature, structure, or characteristic. Furthermore, when a specific feature, structure, or characteristic is described in connection with an embodiment, implementing such a feature, structure, or characteristic in conjunction with other embodiments (whether explicitly described or not) should be within the knowledge of those skilled in the art.

[0022] Generally, terms can be understood at least partly from their use in context. For example, depending at least partly on the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in a singular sense, or a combination of features, structures, or characteristics in a plural sense. Additionally, the term "based on" can be understood not necessarily to convey an exclusive set of factors, but rather, alternatively, depending at least partly on the context, to allow for the presence of other factors that are not necessarily explicitly described.

[0023] like Figure 1 As shown, a method for detecting the thickness of the tear film lipid layer includes the following steps: S1, Data Acquisition: Interference image data of the tear film on the ocular surface is acquired through specialized equipment, including a white light source, an interferometer, a camera, and a computer system; S2, Image preprocessing: The acquired interferometric image data is preprocessed, including image denoising, contrast enhancement and normalization. S3, Feature Extraction: Perform feature extraction on the preprocessed interferometric image to extract color, texture and shape features from the image and generate feature vectors. Feature extraction includes color histogram analysis, texture analysis and shape analysis to extract color information, texture patterns and shape features from the interferometric image, respectively. S4, Model Training: Based on the extracted feature vectors, a machine learning algorithm is used to construct and train a tear film lipid layer thickness prediction model; S5, Thickness Calculation: Input the interference image data of the tear film to be detected into the trained prediction model, calculate the thickness of the tear film lipid layer, output the thickness value, and generate a detection report; S6, Result Verification: Verify the calculation results by comparing them with standard sample data and expert evaluation to verify the accuracy of the test results; S7, Clinical Application: Apply the validated model and method to clinical tear film lipid layer thickness detection to assist doctors in the diagnosis and treatment of meibomian gland dysfunction.

[0024] S1 specifically includes: White light source: used to illuminate the tear film on the surface of the eye, providing uniform and stable illumination; Interferometer: Used to generate interference images of the tear film by reflecting white light through the upper and lower interfaces of the tear film lipid layer, producing interference fringes; Camera: Used to capture interferometric images, ensuring high resolution and high quality; Computer system: used to store and process acquired interferometric image data; Example of specific steps: Equipment preparation: Select a suitable white light source (such as an LED light source) to ensure uniform and stable light intensity; calibrate the interferometer to ensure it can accurately generate interference images of the tear film; install the camera and adjust its position and focal length to ensure it can clearly capture interference images; configure the computer system to ensure it has sufficient storage and processing capabilities to store and process large amounts of image data in real time. Patient preparation: Have the patient sit in a comfortable position with their head fixed in place, ensuring their eyes are stable and open, and ensuring the surface of their eyes is clean to avoid any impurities or medication interference; Data acquisition process: A white light source is activated to illuminate the patient's tear film. The position and angle of the light source are adjusted to ensure that the light uniformly covers the entire tear film area. An interference image of the tear film is generated by an interferometer. The interference image is formed by the interference of light reflected from the upper and lower interfaces of the tear film lipid layer. The camera captures the generated interference image in real time. The positions of the camera and the interferometer are kept stable to ensure that the captured image is clear and jitter-free. The captured interference image data is transmitted to the computer system, which performs preliminary processing and storage of the image data.

[0025] S2 specifically includes: S21, Image Denoising: Gaussian filtering is used to denoise the acquired interference image data; Gaussian filtering smooths images and reduces the impact of noise by convolving with a Gaussian kernel. The Gaussian kernel is a weight matrix whose elements are calculated using a two-dimensional Gaussian function, which is calculated as follows: ; For Gaussian filter at position The value at that location, The standard deviation controls the smoothness of the Gaussian filter; the larger the value, the smoother the filtering effect. These are the position coordinates relative to the center of the filter; S22, Enhance contrast: Histogram equalization is used to enhance the contrast of the acquired interferometric image data; Histogram equalization increases the global contrast of an image by adjusting the distribution of its grayscale values. It is calculated as follows: ; in, For the new grayscale value, This refers to the number of gray levels (usually 256, corresponding to gray levels 0-255). The total number of pixels in the image. For the first The pixel probability of a grayscale level is calculated by dividing the number of pixels at that grayscale level by the total number of pixels. The current grayscale level, Let be the cumulative distribution function (CDF), representing the distribution from 0 to the nth. The cumulative probability of gray levels; S23, Standardization Processing: The acquired interferometric image data is standardized using image standardization methods; Image standardization involves subtracting the mean from the pixel values ​​of an image and dividing by the standard deviation to make the image data conform to a standard normal distribution. The standardization formula is as follows: ; in, These are the standardized pixel values. These are the original pixel values. This is the mean of the image, that is, the average value of all pixel values. This represents the standard deviation of the image, which is the standard deviation of all pixel values.

[0026] S3 specifically includes: S31, Color Histogram Analysis: Color histogram analysis describes the color features of an image by statistically analyzing the pixel value distribution of each color channel. The specific steps are as follows: (1) Color space conversion: Convert the image from RGB color space to HSV (hue-saturation-lightness) color space, because HSV color space is more in line with human visual perception; (2) Histogram calculation: Calculate the histograms of the three channels of HSV respectively, and statistically analyze the frequency distribution of pixel values ​​in the histogram of each channel; (3) Histogram normalization: Normalize the histogram so that the sum of the histograms is 1, which facilitates feature comparison; For color channels (such as H, S, or V channels), its histogram The normalization formula is: ,in, Indicates color channel The Middle The frequency of each gray level Indicates the first Number of pixels in each grayscale level It is a color channel Total number of pixels; S32, Texture Analysis: Texture analysis extracts the texture features of an image by describing the spatial distribution relationship between pixels. The specific steps are as follows: (1) Gray level processing: Convert the image to grayscale and reduce the number of gray levels (e.g., reduce the number of gray levels to 16 or 32) to simplify calculations; (2) Constructing the gray-level co-occurrence matrix: Calculate the gray-level co-occurrence relationship between each pixel in the image and its neighboring pixels, and construct the gray-level co-occurrence matrix; (3) Calculate texture features: Extract statistical features from the gray-level co-occurrence matrix, such as contrast, homogeneity, energy and entropy; Contrast ratio is calculated as follows: Contrast ,in, It is the position in the gray-level co-occurrence matrix. The value, It refers to grayscale levels; Homogeneity is calculated as follows: Homogeneity ; Energy calculation is as follows: ; Entropy is calculated as follows: Entropy ; S33, Shape Analysis: Shape analysis extracts shape features by describing the geometric shape features of objects in an image. The specific steps are as follows: Edge detection: Using edge detection algorithms (such as Canny edge detection) to extract edge contours in an image; Fourier transform: Convert the coordinates of the edge contour into complex form, perform a Fourier transform, and obtain the Fourier descriptor; Feature extraction: Select the first few Fourier descriptors as shape features; Fourier descriptor: a sequence of edge contour points Represent it as a complex number The Fourier transform formula is: ,in, It is the first A Fourier description, It is the total number of edge points; S34, Feature Vector Generation: Combine the extracted color, texture, and shape features into a feature vector; The feature vector will be used as input data for subsequent tear film lipid layer thickness prediction models. By integrating color, texture and shape features, it can comprehensively describe the information in the interference image, improving the accuracy and robustness of the prediction.

[0027] S4 specifically includes: S41, Data Preparation: Combine the extracted color, texture and shape features into a feature vector, and label the corresponding tear film lipid layer thickness values ​​to form a dataset; S42, Select machine learning algorithm: Use convolutional neural network to build tear film lipid layer thickness prediction model. Convolutional neural network has significant advantages in processing image data and can automatically extract and learn the hierarchical features of the image. A convolutional neural network is used to construct a tear film lipid layer thickness prediction model, including: Input layer: Inputs the preprocessed feature vectors; Convolutional layers: extract local features of an image; Pooling layer: Reduces data dimensionality and extracts key features; Fully connected layer: Flatten the features extracted by the convolutional and pooling layers and connect them to the fully connected layer for final thickness prediction; Output layer: Outputs the thickness value of the tear film lipid layer; S43, Model Training: Train the convolutional neural network model using the dataset. The specific steps are as follows: (1) Data partitioning: Divide the dataset into training set, validation set and test set, with a common ratio of 8:1:1; (2) Hyperparameter settings: Set hyperparameters such as learning rate, batch size, and number of iterations to optimize the model training process; (3) Loss function: The mean squared error is chosen as the loss function. The formula for calculating the mean squared error is: ; in, It is a predicted value. It is the actual value. It is the number of samples; (4) Optimization algorithm: Select an optimization algorithm, such as the Adam optimizer, and update the model parameters; (5) Model validation: Use the validation set to evaluate the performance of the model, and use the test set to test the trained model to evaluate the generalization ability and prediction accuracy of the model.

[0028] S5 specifically includes: S51, Preprocessing: The interference image data of the tear film to be detected is preprocessed to ensure that the image data quality meets the requirements of the prediction model. The preprocessing steps include denoising, contrast enhancement and normalization. S52, Feature Extraction: Extract color, texture, and shape features from the preprocessed interferometric image to generate feature vectors. The features include color histogram analysis, gray-level co-occurrence matrix texture analysis, and Fourier descriptor shape analysis. S53, Calculate the tear film lipid layer thickness: Input the generated feature vector into the trained tear film lipid layer thickness prediction model, and the model outputs the predicted tear film lipid layer thickness based on the input feature vector. S54, Generate Detection Report: Based on the predicted thickness value, a detection report is generated. The report includes tear film lipid layer thickness, detection time, and relevant diagnostic suggestions. The specific steps are as follows: Record the testing time: Record the date and time of the test; Output thickness value: Record the thickness value of the tear film lipid layer (e.g., 45nm); Provide diagnostic recommendations: Based on the thickness values, provide relevant diagnostic recommendations, such as assessing the health of the tear film, recommending further examination of meibomian gland function, and providing appropriate treatment options; Through the above steps, the interference image of the tear film to be tested can be effectively input into the trained prediction model to calculate the thickness of the tear film lipid layer and generate a detailed test report, providing doctors with valuable diagnostic information and treatment suggestions.

[0029] S6 specifically includes: S61, Standard Sample Data: Prepare standard sample data. Standard sample data are sample data for which the thickness of the tear film lipid layer is known in advance. They are usually measured by experts using other precise methods (such as clinical assessment or high-precision experimental equipment). S62, Prediction Model Output: Calculate the tear film lipid layer thickness of the test sample using the trained tear film lipid layer thickness prediction model, and record the predicted value for each sample. S63, Result Comparison and Analysis: Compare and analyze the model's prediction results with standard sample data, calculate the error and statistical indicators between the prediction results and the standard values, and evaluate the accuracy of the model; The mean absolute error (MAE) is calculated as follows: MAE ; in, It is a predicted value. It is the true value of the standard sample. It is the number of samples; Statistical indicators: Further evaluation of the model's predictive performance is achieved by calculating other statistical indicators, such as the coefficient of determination. The calculation is as follows: ; in, It is the average of the standard sample. It is a predicted value. It is the actual value; S64, Expert Evaluation: Invite ophthalmologists to evaluate the prediction results of the samples. The experts review and confirm the prediction results based on their clinical experience and professional knowledge, and judge whether the model prediction results are reasonable. S65, Model Adjustment and Optimization: Based on feedback from error analysis and expert evaluation, the tear film lipid layer thickness prediction model is adjusted and optimized to improve its accuracy and robustness. Adjustment methods include retraining the model, increasing training data, optimizing feature extraction methods, and adjusting model parameters. The above steps can effectively verify the prediction results of the tear film lipid layer thickness detection model, ensuring its accuracy and reliability. By comparing standard sample data and conducting expert evaluation, the model's shortcomings can be identified and corrected in a timely manner, thereby improving the model's practical application value.

[0030] S7 specifically includes: S71, Deployment: Integrate the validated models and detection methods into ophthalmic diagnostic and treatment equipment in hospitals or clinics, ensuring that the system can operate stably and is compatible with other diagnostic and treatment equipment, and provide operation training to medical staff to ensure that doctors and technicians are proficient in using the equipment; S72, Application: During patient examination, a special device is used to acquire different tear film interference images. The acquired interference images are preprocessed, and color, texture, and shape features are extracted from the preprocessed interference images to generate feature vectors. The feature vectors are then input into a trained tear film lipid layer thickness prediction model, and the model outputs the thickness value of the tear film lipid layer.

[0031] S7 further includes: S73, Report Generation: Generates patient test reports, including tear film lipid layer thickness, test time, and related diagnostic recommendations; S74, Diagnosis and Treatment: Based on test reports and clinical information, doctors conduct a comprehensive assessment of the patient's meibomian gland function and provide personalized treatment plans, such as hot compresses, meibomian gland massage, and drug therapy. The tear film lipid layer thickness data provided can help doctors more accurately determine the patient's tear film health status. By applying the above steps to the validated tear film lipid layer thickness detection model and method in clinical practice, doctors can effectively assist in the diagnosis and treatment of meibomian gland dysfunction.

[0032] like Figure 2 As shown, a tear film lipid layer thickness detection system is used to implement the above-mentioned tear film lipid layer thickness detection method, and includes the following modules: Data acquisition module: used to acquire interference image data of the tear film on the ocular surface using dedicated equipment, which includes a white light source, an interferometer, a high-definition camera, and a computer system; Image preprocessing module: used to preprocess the acquired interferometric image data, including image denoising, contrast enhancement and normalization. Feature extraction module: used to extract features from the preprocessed interference image, extracting color, texture and shape features from the image and generating feature vectors. The feature extraction includes color histogram analysis, texture analysis and shape analysis, which respectively extract color information, texture patterns and shape features from the interference image. Model training module: used to build and train a tear film lipid layer thickness prediction model based on the extracted feature vectors and using machine learning algorithms; Thickness calculation module: This module is used to input the interference image of the tear film to be detected into the trained prediction model, calculate the thickness of the tear film lipid layer, output the thickness value, and generate a detection report. Result verification module: Used to verify the calculation results by comparing them with standard sample data and expert evaluations to verify the accuracy of the test results; Clinical Application Module: This module is used to apply validated models and methods to clinical tear film lipid layer thickness detection, assisting doctors in the diagnosis and treatment of meibomian gland dysfunction.

[0033] This invention encompasses any substitutions, modifications, equivalent methods, and solutions made within the spirit and scope of this invention. To provide the public with a thorough understanding of this invention, specific details are described in detail in the following preferred embodiments; however, those skilled in the art will fully understand the invention even without these details. Furthermore, to avoid unnecessary misunderstanding of the essence of this invention, well-known methods, processes, procedures, components, and circuits are not described in detail.

[0034] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for detecting the thickness of the tear film lipid layer, characterized in that, Includes the following steps: S1, Data Acquisition: Interference image data of the tear film on the ocular surface is acquired through a dedicated device, which includes a white light source, an interferometer, a camera, and a computer system; S2, Image preprocessing: The acquired interferometric image data is preprocessed, including image denoising, contrast enhancement and normalization. S3, Feature Extraction: Perform feature extraction on the preprocessed interferometric image to extract color, texture and shape features from the image and generate feature vectors. The feature extraction includes color histogram analysis, texture analysis and shape analysis to extract color information, texture patterns and shape features from the interferometric image, respectively. S4, Model Training: Based on the extracted feature vectors, a machine learning algorithm is used to construct and train a tear film lipid layer thickness prediction model; S5, Thickness Calculation: Input the interference image data of the tear film to be detected into the trained prediction model, calculate the thickness of the tear film lipid layer, output the thickness value, and generate a detection report; S6, Result Verification: Verify the calculation results by comparing them with standard sample data and expert evaluation to verify the accuracy of the test results; S7, Clinical Application: Apply the validated model and method to clinical tear film lipid layer thickness detection to assist doctors in the diagnosis and treatment of meibomian gland dysfunction.

2. The method for detecting tear film lipid layer thickness according to claim 1, characterized in that, S1 specifically includes: White light source: used to illuminate the tear film on the surface of the eye, providing uniform and stable illumination; Interferometer: Used to generate interference images of the tear film by reflecting white light through the upper and lower interfaces of the tear film lipid layer, producing interference fringes; Camera: Used to capture interference images; Computer system: Used to store and process acquired interferometric image data.

3. The method for detecting tear film lipid layer thickness according to claim 2, characterized in that, S2 specifically includes: S21, Image Denoising: Gaussian filtering is used to denoise the acquired interference image data; S22, Enhance contrast: Histogram equalization is used to enhance the contrast of the acquired interferometric image data; S23, Standardization Processing: The acquired interferometric image data is standardized using image standardization methods.

4. The method for detecting tear film lipid layer thickness according to claim 3, characterized in that, S3 specifically includes: S31, Color Histogram Analysis: Color histogram analysis describes the color characteristics of an image by statistically analyzing the pixel value distribution of each color channel in the image. S32, Texture Analysis: Texture analysis extracts the texture features of an image by describing the spatial distribution relationship between pixels in the image; S33, Shape Analysis: Shape analysis extracts shape features by describing the geometric shape features of objects in an image; S34, Feature Vector Generation: Combine the extracted color, texture and shape features into a feature vector.

5. The method for detecting tear film lipid layer thickness according to claim 4, characterized in that, S4 specifically includes: S41, Data preparation: Combine the extracted color, texture and shape features into a feature vector, and label the corresponding tear film lipid layer thickness values ​​to form a dataset; S42, Select machine learning algorithm: Use convolutional neural network to build tear film lipid layer thickness prediction model; S43, Model Training: Train the convolutional neural network model using the dataset.

6. The method for detecting tear film lipid layer thickness according to claim 5, characterized in that, S5 specifically includes: S51, Preprocessing: The interference image data of the tear film to be detected is preprocessed. The preprocessing steps include noise reduction, contrast enhancement and normalization. S52, Feature Extraction: Extract color, texture, and shape features from the preprocessed interference image to generate feature vectors; S53, Calculate the tear film lipid layer thickness: Input the generated feature vector into the trained tear film lipid layer thickness prediction model, and the model outputs the predicted tear film lipid layer thickness based on the input feature vector. S54, Generate Test Report: Based on the predicted thickness value, generate a test report, which includes the tear film lipid layer thickness, test time, and related diagnostic recommendations.

7. The method for detecting tear film lipid layer thickness according to claim 6, characterized in that, S6 specifically includes: S61, Standard Sample Data: Prepare standard sample data, which are sample data for which the thickness of the tear film lipid layer is known in advance; S62, Prediction Model Output: Calculate the tear film lipid layer thickness of the test sample using the trained tear film lipid layer thickness prediction model, and record the predicted value for each sample. S63, Result Comparison and Analysis: Compare and analyze the model's prediction results with standard sample data, calculate the error and statistical indicators between the prediction results and the standard values, and evaluate the accuracy of the model; S64, Expert Evaluation: Invite ophthalmologists to evaluate the prediction results of the sample; S65, Model Adjustment and Optimization: Based on feedback from error analysis and expert evaluation, the tear film lipid layer thickness prediction model is adjusted and optimized to improve its accuracy and robustness.

8. The method for detecting tear film lipid layer thickness according to claim 7, characterized in that, Specifically, S7 includes: S71, Deployment: Integrating validated models and detection methods into ophthalmic diagnostic and treatment equipment in hospitals or clinics; S72, Application: During patient examination, a special device is used to acquire different tear film interference images. The acquired interference images are preprocessed, and color, texture, and shape features are extracted from the preprocessed interference images to generate feature vectors. The feature vectors are then input into a trained tear film lipid layer thickness prediction model, and the model outputs the thickness value of the tear film lipid layer.

9. The method for detecting tear film lipid layer thickness according to claim 8, characterized in that, S7 further includes: S73, Report Generation: Generates patient test reports, including tear film lipid layer thickness, test time, and related diagnostic recommendations; S74, Diagnosis and Treatment: Based on test reports and clinical information, doctors conduct a comprehensive assessment of the patient's meibomian gland function.

10. A tear film lipid layer thickness detection system, used to implement the tear film lipid layer thickness detection method as described in any one of claims 1-9, characterized in that, Includes the following modules: Data acquisition module: used to acquire interference image data of the tear film on the ocular surface using dedicated equipment, which includes a white light source, an interferometer, a high-definition camera, and a computer system; Image preprocessing module: used to preprocess the acquired interferometric image data, including image denoising, contrast enhancement and normalization. Feature extraction module: used to extract features from the preprocessed interference image, extracting color, texture and shape features from the image and generating feature vectors. The feature extraction includes color histogram analysis, texture analysis and shape analysis, which respectively extract color information, texture patterns and shape features from the interference image. Model training module: used to build and train a tear film lipid layer thickness prediction model based on the extracted feature vectors and using machine learning algorithms; Thickness calculation module: This module is used to input the interference image of the tear film to be detected into the trained prediction model, calculate the thickness of the tear film lipid layer, output the thickness value, and generate a detection report. Result verification module: Used to verify the calculation results by comparing them with standard sample data and expert evaluation to verify the accuracy of the test results; Clinical Application Module: This module is used to apply validated models and methods to clinical tear film lipid layer thickness detection, assisting doctors in the diagnosis and treatment of meibomian gland dysfunction.

Images