An automatic measurement method for the minimum diameter of a macular hole in an eye fundus based on an OCT image

Abstract

The application discloses an automatic measurement method for the minimum diameter of a fundus macular hole based on an OCT image, and particularly relates to the technical field of medical image processing, data preparation, acquisition of a public fundus macular hole OCT dataset, and image data cropping, so that only part of a macular hole lesion area is reserved in the image, and preparation is made for image processing; image processing, including image denoising, image binarization, image morphological operation and edge detection; firstly, the image denoising adopts a superimposed mode of Gaussian filtering in linear filtering and median filtering in nonlinear filtering, then a global threshold is automatically sought by using a triangle method, and the denoised image is binarized according to the threshold, so that an image containing only black and white is generated; the cyst cavity area in the image is removed through image morphological operation processing; and finally, edge detection is performed.

Description

Technical Field

[0001] This invention relates to the field of medical image processing technology, specifically to an automatic method for measuring the minimum diameter of the macular hole in the fundus based on OCT images. Background Technology

[0002] Macular hole (MH) refers to a localized, full-thickness defect of the neuroepithelial layer of the macula. Based on its etiology, it can be divided into idiopathic and secondary macular holes. The cause of idiopathic MH is unknown, while secondary MH is often secondary to ocular trauma, macular degeneration, high myopia, and other diseases. Idiopathic full-thickness macular hole (IFTMH) is a complete loss of the retina from the internal limiting membrane to the retinal pigment epithelium in the fovea. It is one of the most common fundus diseases in ophthalmology. The main symptoms include decreased visual acuity, distorted vision, central scotoma, and visual field defects, which severely impact the patient's quality of life. Vitrectomy combined with internal limiting membrane peeling is the surgical treatment for macular holes. The anatomical closure rate after MH surgery can reach over 90%, but postoperative visual recovery still varies from person to person. With the advancement of ophthalmic examination equipment and in-depth research into the pathogenesis, its treatment methods are becoming more diversified and precise.

[0003] The size of the macular hole is defined by the minimum diameter of the full-thickness defect. Clinically, doctors measure the size of the macular hole using OCT (Optical Coherence Tomography) scans. Preoperative macular hole size is correlated with surgical anatomical success rate, and delayed reopening (defined as the reopening of a macular hole closed for 6 months or longer) is also related to the initial preoperative OCT measurement. For giant holes and refractory holes with long disease durations, varying degrees of reopening may occur postoperatively, which has a more severe impact on the patient's vision, and repeat surgery will also face greater risks. Previous studies have shown that the larger the MH (microscopic diameter), the greater the risk of postoperative reopening, significantly affecting clinical indicators such as best corrected visual acuity (BCVA) and visual acuity (VA). According to the classification of macular hole development stages by the International Vitreomacular Traction Research Group, the minimum diameter of the macular hole is an important indicator for clinical diagnosis; therefore, accurate measurement of the macular hole size is of great significance.

[0004] Clinically, doctors typically use embedded manual calipers in ophthalmology software to manually measure the size of macular holes (MH) and classify and stage them based on the measurement results. However, manual measurement is inefficient, inaccurate, and highly dependent on the doctor's personal experience. Secondly, the number of retinal specialists in my country is far from meeting current public health needs, and high-quality ophthalmological resources are scarce. Manual measurement cannot meet the needs of the numerous patients in areas with underdeveloped medical resources.

[0005] Currently, research on the automatic measurement of macular holes is limited. Because OCT images acquired by devices typically have low contrast and high noise, and the edges of the macular hole are often blurred, the task of automatically measuring the size of the macular hole is extremely challenging. In 2023, a method based on the YOLOv7 algorithm was developed for macular hole detection. This method applies Otsu thresholding to the pixels within the detected bounding box to binarize the region, and then draws eight equidistant horizontal lines within this region. The size of the macular hole is calculated based on the number of pixels with a horizontal value of 0 on the shortest of the eight lines. However, this method can only roughly measure the minimum diameter of the macular hole in the horizontal direction, and the measured value does not reflect the true value of the minimum diameter of the macular hole. Summary of the Invention

[0006] Therefore, this invention provides an automatic measurement method for the minimum diameter of the macular hole in the fundus based on OCT images, in order to solve the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution: an automatic measurement method for the minimum diameter of the macular hole in the fundus based on OCT images, comprising:

[0008] Data preparation: Obtain publicly available OCT datasets of macular hole in the fundus and crop the image data so that only a portion of the macular hole lesion area is retained in the image, minimizing interference from other parts of the image, in preparation for image processing;

[0009] Image processing includes image denoising, image binarization, image morphological operations, and edge detection. First, image denoising employs a combination of Gaussian filtering (linear filtering) and median filtering (nonlinear filtering). Then, a global threshold is automatically calculated using the triangle method. Based on this threshold, the denoised image is binarized to generate an image containing only black and white. Next, image morphological operations are performed to remove cystic regions from the image. Finally, edge detection is performed to identify strong and weak edges in the image.

[0010] Automatic measurement involves automatically finding two contour lines in an image, automatically identifying the closest position between the two contour lines, and calculating the minimum Euclidean distance between the two contours.

[0011] Preferably, the Gaussian filtering step in linear filtering includes: a two-dimensional Gaussian function as shown in formula (1),

[0012] The Gaussian kernel generated using formula (1) will filter the image in the x and y directions;

[0013]

[0014] G(x, y) represents the value of the Gaussian function at the point (x, y); σ is the standard deviation of the Gaussian filter, which determines the width of the Gaussian function in each direction; the larger the standard deviation, the wider and smoother the Gaussian function spreads; x and y are independent variables, representing the position of the point in the image.

[0015] Preferably, the specific steps of automatically finding the global threshold using the triangle method include: first, calculating the gray intensity histogram of the image, assuming that the maximum peak of the histogram is on the side closest to the brightest, then obtaining the maximum straight-line distance through triangles, and determining the threshold based on the gray level of the histogram corresponding to the maximum straight-line distance.

[0016] Preferably, morphological operations include closing, opening, erosion, filling, and dilation operations; combining image erosion and image dilation to achieve image closing and opening operations;

[0017] Among them, the closing operation is an operation that first dilates and then erodes the image, as shown in formula (2);

[0018] Closing(A)=Erosion(Dilation(A)) (2);

[0019] A represents the input binary image or the target object of the morphological operation; Erosion represents erosion, one of the morphological operations; Dilation represents dilation, one of the morphological operations.

[0020] The closing operation first dilates and then erodes the image, removing small holes in the connected components, connecting adjacent connected components, and smoothing the edges of the contour; while the opening operation is an operation that first erodes and then dilates the image, as shown in formula (3).

[0021] Opening(A)=Dilation(Erosion(A)) (3);

[0022] The opening operation first erodes and then dilates the image to remove noise, eliminate smaller connected regions, and retain larger connected regions, smoothing the boundaries of connected regions without significantly changing their areas. After performing morphological operations on the binarized image, smaller closed black areas are filled to eliminate cystic structures.

[0023] Preferably, the edge detection process consists of the following steps:

[0024] Step 1: Image noise reduction: Use Gaussian filtering to smooth the image and reduce noise in the image. Generally, a 5*5 Gaussian filter as shown in formula (4) is used.

[0025]

[0026] Step 2: Calculate the image gradient: Calculate the gradient direction and magnitude of the grayscale value of each pixel in the image; First, detect the edges in the x and y directions of the image using the Sobel operator, and then calculate the gradient direction θ and magnitude G using formulas (5) and (6), respectively, where I x I represents the gray level of a pixel in the x-direction. y Represents the gray level of a pixel in the y-direction; the Sobel operator effectively extracts image edges using the discrete differentiation method;

[0027]

[0028] Step 3: Non-maximum suppression: Non-maximum suppression algorithm is used to eliminate stray responses caused by edge detection. The gray-level change area is usually concentrated. The pixel with the largest gray-level gradient in the gradient direction in the local range is retained and the others are not retained. A large number of points can be removed, and the edge with multiple pixels wide can be turned into a single pixel wide edge.

[0029] Step 4: Dual Threshold Filtering: After non-maximum suppression, there are still many possible edge points. The dual threshold method is used to divide strong edges into weak edges. Pixels with gray-level gradients greater than the high threshold are retained as strong edges, and pixels with gray-level gradients less than the low threshold are removed. Pixels with gray-level gradients between the high and low thresholds are filtered as weak edges. The weak edge is then searched in the neighborhood of the weak edge. If a strong edge exists, the weak edge is retained; otherwise, it is deleted, thus completing the edge detection.

[0030] Preferably, the Euclidean distance is calculated as shown in formula (8).

[0031]

[0032] Where x1 and y1 represent the pixel coordinates of the first point, x2 and y2 represent the pixel coordinates of the second point, and d represents the Euclidean distance between the two pixels.

[0033] The present invention has the following advantages:

[0034] This invention combines image denoising, image binarization, image morphological operations, and edge detection algorithms. The algorithm exhibits good robustness and innovation, enabling efficient and accurate automatic measurement of the macular hole diameter, thus alleviating significant clinical burden. Combined with the automatic measurement results, doctors can develop more comprehensive treatment plans for patients. This invention addresses the current limitation of artificial intelligence technology in the treatment of macular hole diseases, demonstrating excellent detection results for the contour edges of macular holes on the vast majority of macular hole OCT data. It is of great significance for promoting the development of precision medicine, intelligent medicine, and telemedicine. Attached Figure Description

[0035] Figure 1 This is a flowchart of the automatic measurement process for the minimum diameter of macular hole provided in this embodiment;

[0036] Figure 2 The OCT image of the macular hole provided in this embodiment;

[0037] Figure 3 This is a schematic diagram illustrating the principle of automatically finding the global threshold using the triangle method provided in this embodiment;

[0038] Figure 4 This is a schematic diagram illustrating the image erosion principle provided in this embodiment;

[0039] Figure 5 This is a schematic diagram illustrating the principle of image dilation provided in this embodiment;

[0040] Figure 6 This is a schematic diagram illustrating the principle of the closing operation provided in this embodiment;

[0041] Figure 7 This is a schematic diagram illustrating the principle of the opening operation provided in this embodiment;

[0042] Figures 8(a) and 8(b) are the binarized image and the image after morphological operation provided in this embodiment, respectively;

[0043] Figures 9(a) and 9(b) show the images after edge detection and after automatic distance measurement provided in this embodiment;

[0044] Figure 10 This is a flowchart illustrating the implementation of automatic measurement of macular holes in this embodiment;

[0045] Figure 11 This is an example diagram of the automatic measurement of macular holes provided in this embodiment. Detailed Implementation

[0046] The following specific embodiments illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0047] Refer to the instruction manual Figure 1 This embodiment of an automatic measurement method for the minimum diameter of the macular hole in the fundus based on OCT images includes:

[0048] 1. Data Preparation

[0049] The main task of this step is to obtain publicly available OCT datasets of macular hole lesions and crop the image data so that only a portion of the macular hole lesion area is retained in the image, minimizing interference from other parts of the image and preparing for image processing.

[0050] like Figure 2 The image shown is an OCT image of a macular hole. The lesion area of ​​the macular hole appears as a low-reflectance region consistent with the vitreous cavity reflection in the OCT B-scan. Its base is an exposed, highly reflective retinal pigment epithelial (RPE) layer, flanked by well-defined, highly reflective retinal tissue. Sometimes, completely or partially detached tissue formed by traction from the internal vitreous membrane can be seen above it. Macular cystoid edema is very common in macular hole cases, appearing in the OCT B-scan as uniformly low-reflective cystic structures of varying sizes and high-reflectance intervals between different retinal tissue layers. Cropping the image data avoids interference from the RPE layer, cystic structures, excessive image noise, and other tissues in subsequent image processing. This step lays the foundation for edge extraction and automatic measurement of the macular hole.

[0051] 2. Image Processing

[0052] The main content of this step is to construct the image processing algorithm, which can be divided into the following steps:

[0053] 2.1 Image Denoising

[0054] The original OCT image contains a lot of noise, which can significantly affect the subsequent image processing results. Choosing a suitable denoising method is crucial for the final image processing outcome. This step aims to preprocess the image by using a combination of linear and nonlinear filtering for denoising. Through experimental comparison, it was found that selecting Gaussian filtering in linear filtering and median filtering in nonlinear filtering yields the best preprocessing results.

[0055] Linear filtering results in a linear combination of all pixel values. Gaussian noise is easily introduced during image acquisition, and Gaussian filtering effectively removes this noise. The essence of Gaussian filtering is to perform a convolution operation on the image using a Gaussian kernel generated by a Gaussian function. The two-dimensional Gaussian function is shown in formula (1). The Gaussian kernel generated by the function filters the image in the x and y directions. Here, σ is the standard deviation of the Gaussian filter. The larger σ is, the wider the bandwidth of the Gaussian filter, resulting in a better smoothing effect on the image and making the image more blurred. It is necessary to select an appropriate Gaussian kernel size and σ value to ensure that the image edges are not excessively blurred while reducing noise.

[0056]

[0057] G(x,y) represents the value of the Gaussian function at point (x,y); σ is the standard deviation of the Gaussian filter, which determines the width of the Gaussian function in each direction; the larger the standard deviation, the wider and smoother the Gaussian function spreads; x and y are independent variables, representing the position of the point in the image.

[0058] The result of nonlinear filtering is not obtained by linearly combining the pixel values ​​within the filter; its operation involves logical operations and sorting. Noise after linear filtering is not eliminated but exists in a smoother form. Using nonlinear filtering, which filters out noise through logical judgment, can achieve better image processing results. Median filtering replaces the pixel value at the center of the filter with the median value of all pixels within the filter range. It is a nonlinear signal processing method based on sorting statistics theory and can effectively suppress noise. Median filtering preserves the edge information of the image very well, but it requires choosing an appropriate filter size to avoid excessive blurring of image details after denoising.

[0059] 2.2 Image Binarization

[0060] Binarizing the denoised image generates an image containing only black and white. Binarization lays the foundation for subsequent image edge extraction. The chosen grayscale threshold directly affects the binarization effect. Based on experimental comparisons, a global threshold is automatically sought using the triangle method. Binarizing the image using this threshold yields the optimal binarized image. This method uses a purely geometric approach to find the optimal threshold. First, the image's grayscale intensity histogram is calculated. Assuming the maximum peak of the histogram is near the brightest side, the maximum straight-line distance is obtained using triangles. The threshold is then determined based on the histogram grayscale level corresponding to the maximum straight-line distance. The principle of this method is as follows: Figure 3 As shown.

[0061] 2.3 Image Morphological Operations

[0062] Since macular cystoid edema is very common in macular hole cases, corresponding cystic structures can be observed in many OCT images. In binarized images, the cystic structures appear as closed black areas. This region affects the extraction of the macular hole's edge. Through image morphological operations, image components meaningful for expressing and describing the shape and location of the region can be extracted from the image, enabling subsequent recognition work to capture the most essential shape characteristics of the object. Therefore, performing morphological operations on binarized images to remove cystic regions is of great significance for subsequent image processing.

[0063] By performing closing, opening, erosion, filling, and dilation operations sequentially on a binarized image, cystic regions in the image can be effectively removed.

[0064] Image erosion and dilation are fundamental morphological operations, requiring a template matrix called a structuring element. Commonly used structuring elements include cross-shaped, rectangular, and elliptical elements; this example uses a 3x3 rectangular structuring element for image processing. A connected component in an image is a region composed of pixels with the same pixel value and adjacent positions. Erosion can remove smaller connected components caused by noise. The principle of image erosion is as follows... Figure 4 As shown, the dilation operation is the opposite of the erosion operation, ensuring that the area of ​​the main region does not shrink. The principle of image dilation is as follows: Figure 5 As shown.

[0065] Combining image erosion and image dilation allows for the implementation of image closing and opening operations. Closing is an operation that involves first dilating and then eroding the image, as shown in formula (2). Closing, by first dilating and then eroding, can remove small holes within connected components, connect adjacent connected components, and smooth the edges of contours. The principle of closing operation is as follows: Figure 6 As shown in equation (3). The opening operation is an operation that erodes and then dilates an image, which can remove noise, eliminate smaller connected components, and retain larger connected components. It smooths the boundaries of connected components without significantly changing the area of ​​larger connected components. The principle of the opening operation is as follows: Figure 7 As shown in Figure 8(a), after performing the above morphological operations on the binarized image, it is necessary to fill the smaller closed black areas to eliminate the cystic structure. The binarized image is shown in Figure 8(a), and the image after morphological operations is shown in Figure 8(b).

[0066] Closing(A)=Erosion(Dilation(A))(2)

[0067] A represents the input binary image or the target object of the morphological operation; Erosion represents erosion, one of the morphological operations; Dilation represents dilation, one of the morphological operations.

[0068] Opening(A)=Dilation(Erosion(A))(3)

[0069] 2.4 Edge Detection

[0070] Image edges refer to regions in an image where the grayscale value of pixels changes abruptly. The Canny algorithm is one of the best and most commonly used edge detection algorithms. It is less susceptible to noise and can identify both strong and weak edges in an image, playing a crucial role in edge detection applications. The algorithm's detection process consists of the following four steps:

[0071] Image noise reduction. Noise is where grayscale changes are large and can easily be identified as false edges. Therefore, Gaussian filtering is used to smooth the image and reduce noise. Generally, a 5*5 Gaussian filter as shown in formula (4) is used.

[0072]

[0073] Calculate the image gradient and detect possible edges. This step requires calculating the gradient direction and magnitude of the grayscale value of each pixel in the image. First, the Sobel operator is used to detect edges in the x and y directions of the image, respectively. Then, the gradient direction θ and magnitude G are calculated using formulas (5) and (6), where I x I represents the gray level of a pixel in the x-direction. y This represents the gray level of a pixel in the y-direction. The Sobel operator can effectively extract image edges using the discrete differentiation method.

[0074]

[0075] Non-maximum suppression. Non-maximum suppression algorithms can eliminate stray responses from edge detection. Regions of grayscale variation are usually concentrated; by retaining the pixel with the largest grayscale gradient along the gradient direction within a local area and discarding the others, a large portion of the pixels can be removed, transforming edges that are multiple pixels wide into single-pixel wide edges.

[0076] Dual-threshold filtering. After non-maximum suppression, many potential edge points still exist. A dual-threshold method is applied to distinguish between strong and weak edges. Pixels with a gray-level gradient greater than the higher threshold are retained as strong edges, while pixels with a gray-level gradient less than the lower threshold are removed. Pixels with a gray-level gradient between the higher and lower thresholds are filtered as weak edges. A search is performed within the neighborhood of each weak edge. If a strong edge exists, the weak edge is retained; otherwise, it is deleted, completing edge detection.

[0077] This invention combines image denoising, image binarization, image morphological operations, and edge detection algorithms, achieving good detection results for the contour edges of macular holes on the vast majority of macular hole OCT data. The algorithm has good robustness and innovation.

[0078] 3. Automatic measurement

[0079] By clicking on the region of interest with the mouse to crop the OCT image of the macular hole, the algorithm sequentially performs denoising, binarization, morphological operations, and edge detection on the image. After finding the contours of the two macular holes, the algorithm automatically measures the minimum Euclidean distance between the contour lines of the macular holes. The process is as follows: Figure 10 As shown.

[0080] This invention demonstrates good measurement performance on most images in the publicly available macular hole dataset, Macular Hole Retinal OCT. The figure below shows the measurement results for a portion of the data. Figure 11 As shown.

[0081] This invention constructs an automatic measurement algorithm. The algorithm automatically finds two contour lines in an image, locates the closest position between the two contour lines, and calculates the minimum Euclidean distance between the two contours. The Euclidean distance is the straight-line distance between two pixels, and the calculation formula is shown in formula (8), where x and y represent pixel coordinates. The image after edge detection and the image after automatic distance measurement are shown in Figure 9(a) and (b).

[0082]

[0083] Where x1 and y1 represent the pixel coordinates of the first point, x2 and y2 represent the pixel coordinates of the second point, and d represents the Euclidean distance between the two pixels.

[0084] The macula is the concave area at the posterior pole of the retina and is the most sensitive area for visual function. Macular hole (MH) is one of the most common fundus diseases in ophthalmology clinics, and it has an extremely serious impact on patients' vision. For current MH cases, doctors usually make a diagnosis by manually measuring the size of MH in OCT images using the calipers provided in the software. However, the traditional diagnostic method has the following drawbacks: (1) Manual measurement is inefficient and inaccurate, highly dependent on the doctor's personal experience, and there is human error in the measurement process; (2) In reality, the number of fundus specialists is far from meeting the current public health needs, and there are few high-quality ophthalmological medical resources. Manual measurement cannot meet the needs of many patients in areas with underdeveloped medical resources.

[0085] This invention enables automated measurement of the macular hole diameter. The measurement process is efficient and the results are accurate, alleviating significant clinical burden. Combined with the automated measurement results, doctors can develop more comprehensive treatment plans for patients. This invention addresses the current limitations of artificial intelligence technology in the treatment of macular hole diseases and is of great significance for promoting the development of precision medicine, intelligent medicine, and telemedicine.

[0086] Although the present invention has been described in detail above with general descriptions and specific embodiments, modifications or improvements can be made to it, which will be obvious to those skilled in the art. Therefore, all such modifications or improvements made without departing from the spirit of the present invention fall within the scope of protection claimed by the present invention.

Claims

1. An automatic method for measuring the minimum diameter of the macular hole in the fundus based on OCT images, characterized in that: include: Data preparation: Obtain publicly available OCT datasets of macular hole lesions and crop the image data to retain only a portion of the macular hole lesion area in the image, in preparation for image processing; Image processing includes image denoising, image binarization, image morphological operations, and edge detection. First, image denoising uses a combination of Gaussian filtering (linear filtering) and median filtering (nonlinear filtering). Then, the triangulation method is used to automatically find a global threshold. Based on this threshold, the denoised image is binarized to generate an image containing only black and white. Finally, image morphological operations are performed to remove cystic regions from the image. Finally, edge detection is performed to identify strong and weak edges in the image; Automatic measurement: By automatically finding two contour lines in the image, automatically finding the position where the two contour lines are closest and calculating the minimum Euclidean distance between the two contours; The specific steps of automatically finding the global threshold using the triangle method include: first, calculating the gray intensity histogram of the image, assuming that the maximum peak of the histogram is on the side closest to the brightest, then obtaining the maximum straight-line distance through a triangle, and determining the threshold based on the gray level of the histogram corresponding to the maximum straight-line distance; Morphological operations include closing, opening, erosion, filling, and dilation; combining image erosion and dilation enables closing and opening operations. The closing operation is an operation that first dilates and then erodes the image. Closing(A)=Erosion(Dilation(A)) A represents the input binary image or the target object of the morphological operation; Erosion represents erosion, one of the morphological operations; Dilation represents dilation, one of the morphological operations. The opening operation is an operation that first erodes and then dilates an image. Opening(A)=Dilation(Erosion(A)) After performing morphological operations on the binarized image, smaller closed black areas are filled to eliminate cystic structures. The edge detection process consists of the following steps: Step 1: Image Denoising: Smooth the image using a Gaussian filter, using a 5x5 Gaussian filter as shown below; Step 2: Calculate image gradient: Calculate the gradient direction and magnitude of the grayscale value of each pixel in the image; first, detect the edges in the x and y directions of the image using the Sobel operator, and then calculate the gradient direction θ and magnitude G respectively, where I x I represents the gray level of a pixel in the x-direction. y Represents the gray level of a pixel in the y-direction; the Sobel operator effectively extracts image edges using the discrete differentiation method; Step 3: Non-maximum suppression: The non-maximum suppression algorithm is used to eliminate the stray response caused by edge detection; the pixel with the largest gray-level gradient in the gradient direction within the local range is retained, while the others are not retained, thus turning an edge that is multiple pixels wide into an edge that is only one pixel wide. Step 4: Dual threshold screening: After non-maximum suppression, a dual threshold method is used to divide strong edges into weak edges. Pixels with gray-level gradients greater than the high threshold are retained as strong edges, and pixels with gray-level gradients less than the low threshold are removed. Pixels with gray-level gradients between the high and low thresholds are screened as weak edges. The weak edge is then searched in its neighborhood. If a strong edge exists, the weak edge is retained; otherwise, it is deleted, thus completing edge detection.

2. The automatic measurement method for the minimum diameter of the macular hole in the fundus based on OCT images according to claim 1, characterized in that: The Gaussian filtering steps in linear filtering include: a two-dimensional Gaussian function as shown in formula (1), The Gaussian kernel generated using formula (1) will filter the image in the x and y directions; G(x,y) represents the value of the Gaussian function at point (x,y); I is the standard deviation of the Gaussian filter, which determines the width of the Gaussian function in each direction; x and y are independent variables representing the position of the point in the image.

3. The automatic measurement method for the minimum diameter of the macular hole in the fundus based on OCT images according to claim 1, characterized in that: Euclidean distance calculation: Where x1 and y1 represent the pixel coordinates of the first point, x2 and y2 represent the pixel coordinates of the second point, and d represents the Euclidean distance between the two pixels.

Images