Method for Measuring Strabismus Degree of Patients with Rotary Strabismus Based on Fundus Photographs

Abstract

This application relates to the technical field of strabismus auxiliary measurement, specifically to a method for measuring the strabismus degree of patients with rotary strabismus based on fundus photographs. The method includes: collecting fundus images and performing gray-scale processing to obtain rotary strabismus images; obtaining the fundus structure difference degree of each pixel point according to the gray-scale value difference situation between the eight-neighborhood pixel points of each pixel point in the rotary strabismus image; obtaining the number of neighborhood similar structure of each pixel point according to the difference distribution between the dissimilarity degree of the fundus structure of each pixel point and all its eight-neighborhood pixel points; obtaining the adaptive smoothing parameter of each pixel point according to the number of neighborhood similar structure; performing filtering processing on the rotary strabismus image according to the adaptive smoothing parameter of each pixel point to obtain a rotary strabismus measurement image, detecting the circles in the rotary strabismus measurement image to obtain the macula and the optic disc, and calculating the rotary strabismus degree. This application can improve the accuracy of strabismus degree measurement.

Description

Technical Field

[0001] This application relates to the technical field of strabismus auxiliary measurement, and specifically to a method for measuring the strabismus degree of patients with rotary strabismus based on fundus photographs. Background Art

[0002] Rotary strabismus is a complex abnormal eye movement, manifested as the rotational tilt of one or both eyes around their anteroposterior axis to the temporal or nasal side. This condition often coexists with superior and inferior strabismus, mainly due to the abnormal function of the vertical muscles of the eye, such as overactivity or weakness of the superior rectus or inferior rectus muscles. Rotary strabismus not only affects the eye movement function and has a significant impact on visual function, but may also cause discomfort such as headache and nausea. Rotary strabismus is generally detected by measuring the angle between the macula and the lower 1 / 3 of the optic disc. The accuracy of rotary strabismus measurement is mainly limited by the positioning accuracy of the macula and optic disc of the eye. Currently, the positioning of the macula and optic disc of the eye is usually achieved through the processing and recognition of fundus images.

[0003] Although the relative position changes of the macula and optic disc can be judged through fundus photographs, affected by the clarity of the photographs, their positions cannot be accurately determined, and the rotational angle of the eye cannot be accurately quantified, resulting in insufficient sensitivity in cases of minor or slight rotary strabismus and difficulty in directly observing it in the photograph. In order to clearly identify the macula and optic disc, generally, a filtering algorithm is used to denoise the fundus image to enhance the clarity of the image, and the enhanced image is used to identify the macula and optic disc. However, due to the characteristics of the same structures in the patient's fundus image being connected and similar, when the same filtering parameter is used to denoise the entire fundus image, it is easy to reduce the clarity of the image at the edge of the macula and optic disc, affecting the accuracy of strabismus degree measurement. Summary of the Invention

[0004] In order to solve the above technical problems, this application provides a method for measuring the strabismus degree of patients with rotary strabismus based on fundus photographs to solve the existing problems.

[0005] The method for measuring the strabismus degree of patients with rotary strabismus based on fundus photographs in this application adopts the following technical solutions:

[0006] An embodiment of this application provides a method for measuring the strabismus degree of patients with rotary strabismus based on fundus photographs, and this method includes the following steps:

[0007] S1, collect a fundus image and perform grayscale processing to obtain a rotary strabismus image;

[0008] S2. Obtain the neighborhood distribution sequence of each pixel point according to the gray value difference situation between the eight-neighborhood pixel points of each pixel point in the rotary strabismus image; obtain the fundus structure difference degree of each pixel point according to the gray value difference between the front and rear pixel points of all adjacent pixel points in the neighborhood distribution sequence of each pixel point.

[0009] S3. Obtain the number of neighborhood similar structures of each pixel point according to the difference distribution between the fundus structure dissimilarity degree of each pixel point and that of all its eight-neighborhood pixel points; obtain the adaptive smoothing parameter of each pixel point during the filtering process according to the number of neighborhood similar structures of each pixel point.

[0010] S4. Perform filtering processing on the rotary strabismus image according to the adaptive smoothing parameter of each pixel point to obtain a rotary strabismus measurement image, detect the circles in the rotary strabismus measurement image to obtain the macula and the optic disc; measure the strabismus degree according to the positions of the macula and the optic disc.

[0011] Further, the method for obtaining the neighborhood distribution sequence of each pixel point includes:

[0012] Denote any pixel point in the rotary strabismus image as the target point; obtain the pixel point with the minimum gray value among the eight-neighborhood pixel points of the target point, and denote it as the starting point; starting from the starting point, arrange all the eight-neighborhood pixel points in the sorting direction to obtain the neighborhood distribution sequence of the target point.

[0013] Further, the method for obtaining the sorting direction includes:

[0014] Respectively obtain the two pixel points adjacent to the starting point among the eight-neighborhood pixel points, and denote them as the two adjacent points of the starting point; obtain the difference between the gray value of the starting point and that of each of its adjacent points; take the adjacent point with the minimum difference as the direction point, where if the differences of the two adjacent points are the same, choose any one of the adjacent points as the direction point; if the direction point is in the clockwise direction of the starting point, the sorting direction is the clockwise direction; otherwise, the sorting direction is the counterclockwise direction.

[0015] Further, the method for obtaining the fundus structure difference degree of each pixel point includes:

[0016] For the neighborhood distribution sequence of each pixel point, obtain the fundus structure recognition degree between each pair of adjacent pixel points according to the gray value difference between all pixel points before and after each pair of adjacent pixel points in the neighborhood distribution sequence.

[0017] In the neighborhood distribution sequence, denote the previous and subsequent pixel points of the pair of adjacent pixel points with the maximum fundus structure recognition degree as the front structure discrimination point and the rear structure discrimination point respectively.

[0018] Obtain the fundus structure difference degree of each pixel according to the gray value difference between all pixel points before the pre-structure discrimination point and all pixel points after the post-structure discrimination point.

[0019] Further, the method for obtaining the fundus structure recognition degree between each pair of adjacent pixel points includes: Denote the fundus structure recognition degree between the i-th and the (i + 1)-th pixel points in the neighborhood distribution sequence as , ; where , respectively represent the gray values of the j-th and the k-th pixel points in the neighborhood distribution sequence; i represents the serial number of the previous pixel point in the current two adjacent pixel points in the neighborhood distribution sequence; n represents the number of pixel points in the neighborhood distribution sequence; || is the absolute value symbol.

[0020] Further, the method for obtaining the fundus structure difference degree of each pixel according to the gray value difference between all pixel points before the pre-structure discrimination point and all pixel points after the post-structure discrimination point includes:

[0021] Obtain the average gray value of all pixel points from the first pixel point in the neighborhood distribution sequence to the pre-structure discrimination point, and denote it as the first structure feature value; obtain the average gray value of all pixel points from the post-structure discrimination point to the last pixel point in the neighborhood distribution sequence, and denote it as the second structure feature value; respectively obtain the differences between the gray value of each pixel point and the first structure feature value and the second structure feature value, and denote them as the first difference and the second difference; take the minimum value between the first difference and the second difference as the fundus structure difference degree of each pixel point.

[0022] Further, the method for obtaining the number of neighborhood similar structures of each pixel point includes:

[0023] Obtain the segmentation threshold according to the difference situation of the fundus structure difference degrees between all pixel points and their respective eight-neighborhood pixel points;

[0024] Count the number of cases where the fundus structure dissimilarity between each pixel point and all its eight-neighborhood pixel points is less than the segmentation threshold, and denote it as the number of neighborhood similar structures of each pixel point.

[0025] Further, the method for obtaining the segmentation threshold includes:

[0026] Obtain the difference between the fundus structure difference degrees between each pixel point and its respective eight-neighborhood pixel points as the fundus structure dissimilarity between each pixel point and its respective eight-neighborhood pixel points;

[0027] Conduct Otsu threshold analysis on the fundus structure dissimilarities between all pixel points and all their eight-neighborhood pixel points to obtain the segmentation threshold.

[0028] Further, obtaining the adaptive smoothing parameter of each pixel point during the filtering process according to the number of neighborhood similarity structures of each pixel point includes: denoting the adaptive smoothing parameter of the a-th pixel point in the rotation-type strabismus image as , , where in the formula represents a preset initial smoothing parameter; represents a preset adjustment factor; represents the number of neighborhood similarity structures of the a-th pixel point in the rotation-type strabismus image; represents the sigmoid normalization function.

[0029] Further, the method for obtaining the macula and the optic disc includes:

[0030] Substituting the adaptive smoothing parameter of each pixel point in the rotation-type strabismus image into the NLM filtering algorithm to denoise the rotation-type strabismus image, and obtaining a rotation-type strabismus measurement image;

[0031] Obtain the circles in the rotation-type strabismus measurement image; take the inner area of the circle with the second largest diameter as the macula; take the inner area of the circle with the smallest diameter as the optic disc.

[0032] This application has at least the following beneficial effects:

[0033] By analyzing the influence of the connection and similarity characteristics of the same structure in the fundus image on the filtering effect, this application first analyzes the characteristics of different eye structures in the fundus image. According to the certain continuity of the same structure, using the local gray difference of pixel points, the fundus structure difference degree of each pixel point is obtained, which reflects the position of the pixel point in the fundus structure and preliminarily judges the noise situation; further analyzes the similarity of the fundus structure difference degree between the pixel points of the same structure and the pixel points in their neighborhood, and obtains the number of neighborhood similarity structures of each pixel point. The more the number of neighborhood similarity structures, the more obvious the fundus structure characteristics of the pixel point; in order to exclude the interference of noise, the adaptive smoothing parameter of each pixel point is determined according to the number of neighborhood similarity structures of the pixel point, and local filtering is performed on each pixel point in the image according to the adaptive smoothing parameter, which improves the clarity of the image after filtering, can accurately identify the positions of the macula and the optic disc, and makes the strabismus degree measurement result more accurate. Description of the Drawings

[0034] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained according to these drawings without creative efforts.

[0035] Figure 1 The flowchart of the steps of the method for measuring the strabismus degree of patients with rotary strabismus based on fundus photos provided by this application;

[0036] Figure 2 The block diagram for obtaining the difference degree of fundus structures provided by an embodiment of this application. Detailed implementation manners

[0037] Next, the technical solutions in the embodiments of this application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all the embodiments. Without conflict, the embodiments of this application and the technical features in the embodiments can be combined with each other. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in this application without creative efforts shall fall within the protection scope of this application.

[0038] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which this application belongs.

[0039] Next, the specific solution of the method for measuring the strabismus degree of patients with rotary strabismus provided by this application will be specifically described in conjunction with the accompanying drawings.

[0040] A method for measuring the strabismus degree of patients with rotary strabismus based on fundus photos provided by an embodiment of this application. Specifically, the following method for measuring the strabismus degree of patients with rotary strabismus is provided. Please refer to Figure 1 , and this method includes the following steps:

[0041] S1, collect fundus images and perform grayscale processing to obtain rotary strabismus images.

[0042] The impact of rotary strabismus on visual function is mainly reflected in the disruption of visual fusion. Visual fusion refers to the process in which the visual information received by both eyes is integrated into a clear image in the brain. When the rotation angle of the eyes is abnormal, the images seen by both eyes are no longer accurately aligned, resulting in difficult or impossible visual fusion. This fusion disorder not only causes diplopia (seeing repeated images), but may also further trigger a series of systemic symptoms, such as headache, nausea, vomiting, etc., and seriously affect the quality of daily life. Currently, the main methods for measuring the strabismus degree of rotary strabismus are as follows:

[0043] (1)Maddox rod test: The Maddox rod test is a simple and commonly used subjective measurement method. In this test, a Maddox rod containing red parallel lines is used to refract light into a line, and the patient reports the situation of cyclotropia by observing the positional relationship between the red line and the visual target. This method can evaluate the vertical and horizontal deviations of the eyes and can also roughly measure the rotational deviation.

[0044] (2)Hess screen examination: The Hess screen test is a tool commonly used to evaluate eye movement disorders, especially when evaluating strabismus. In this method, the patient is required to sit in front of a screen with a grid and record the responses of each eye to different visual targets through eye movements in different directions. By analyzing the tracking trajectory of the eyes to visual targets, the presence and degree of cyclotropia can be evaluated.

[0045] (3)Synoptophore examination: The synoptophore is a device used to measure the binocular coordination ability. It provides different visual targets to the two eyes respectively, tests the eye movement responses of the patient, and records the changes in eye position. The synoptophore can not only measure horizontal and vertical strabismus but also be used to evaluate cyclotropia and is one of the commonly used examination methods in clinical practice.

[0046] (4)Fundus photography: It is judged according to the relative positions of the macula and the optic disc in the fundus photograph. The center point of the macula is not completely horizontally aligned with the optic disc but is slightly lower than the center of the optic disc and is located in the middle and lower one-third area of the optic disc. If the position of the macula relative to the optic disc is abnormal (such as significant upward or downward displacement of the macula), it may indicate cyclotropia of the eyeball.

[0047] The most widely used method is to locate the macula and the optic disc through fundus photography and measure the strabismus degree based on the abnormal position of the macula relative to the optic disc. This application aims to accurately determine the positions of the macula and the optic disc through fundus photography and then measure the cyclotropia degree of the patient based on this.

[0048] Specifically, a specific implementation scenario of the embodiment of this application is the fundus image filtering scenario during the measurement of the strabismus degree of cyclotropia; in the detection of cyclotropia in patients, it is usually necessary to locate the macula and the optic disc of the eyeball, and measure the strabismus degree by measuring the angle between the macula and the position at the middle and lower 1 / 3 of the optic disc.

[0049] Therefore, in this embodiment, a fundus image is first obtained by a non-mydriatic fundus color camera. The shooting method is as follows: place the patient's mandible on the jaw rest of the detector, and ensure that the forehead is closely attached to the headrest to maintain the stability of the head and avoid any movement during shooting. The obtained original fundus image is an RGB color image. In order to better perform the auxiliary detection of strabismus, in this embodiment, the weighted average method is used to convert the color image into a grayscale image, denoted as the rotational strabismus image. The weighted average method is a well-known technology, and the specific calculation process will not be elaborated here.

[0050] During the process of locating the macula and optic disc of the eyeball based on the rotational strabismus image, the positioning accuracy is limited by the image clarity. During the filtering process of the rotational strabismus image, due to the characteristics of connection and similarity of the same structures in the patient's fundus image, using the same filtering parameter for the entire rotational strabismus image is likely to cause a decrease in the clarity of the image at the edges of the macula and optic disc. Therefore, in this embodiment, the local distribution characteristics of pixel points are used to analyze the similarity characteristics of local structures, and the recognition degree of the fundus structure of pixel points is judged; the filtering parameter is adaptively determined according to the recognition degree of the fundus structure of pixel points.

[0051] S2. According to the gray value difference situation between the eight-neighborhood pixel points of each pixel point in the rotational strabismus image, obtain the neighborhood distribution sequence of each pixel point; according to the gray value difference between the front and back pixel points of all adjacent pixel points in the neighborhood distribution sequence of each pixel point, obtain the fundus structure difference degree of each pixel point.

[0052] For the captured fundus image, it includes the blood vessels, retina, macula of the eyeball, and optic disc of the fundus. The local features of the fundus image have relatively high similarity features. When the pixel point in the rotational strabismus image is a noise point, the pixel value of the noise point has a large difference from its local neighborhood pixel points. When denoising, the denoising intensity should be enhanced at the position of the noise point.

[0053] Obtain the eight-neighborhood pixel points of each pixel point of the rotational strabismus image, and use the mean filling method to complement the insufficient pixel points in the window for the edge pixel points of the image. Since the pixel points of the rotational strabismus image have a relatively high similarity in the same structure; among them, the main structures are blood vessels, retina, macula of the eyeball, and optic disc. There are obvious differences between different structures. For example, there is an obvious gray value difference between the retina and the retinal blood vessels.

[0054] If pixel points are formed by the same structure, the difference between the gray values of pixel points is small. To distinguish the composition of the local neighborhood of pixel points, the eight-neighborhood pixels of each pixel point are sorted. The sorting method is as follows: Denote any pixel point in the rotationally oblique image as the target point; Obtain the pixel point with the smallest gray value among the eight-neighborhood pixel points of the target point, and denote it as the starting point; Starting from the starting point, arrange all the eight-neighborhood pixel points in the sorting direction to obtain the neighborhood distribution sequence of the target point.

[0055] Among them, the method for obtaining the sorting direction is as follows: Respectively obtain the two pixel points adjacent to the starting point among the eight-neighborhood pixel points, and denote them as the two adjacent points of the starting point; Obtain the differences between the gray values of the starting point and its respective adjacent points; Take the adjacent point with the smallest difference as the direction point. Among them, if the differences of the two adjacent points are the same, choose any one of the adjacent points as the direction point; If the direction point is in the clockwise direction of the starting point, the sorting direction is the clockwise direction; Otherwise, the sorting direction is the counterclockwise direction. The two pixel points adjacent to the starting point represent the two pixel points closest to the starting point in distance.

[0056] The direction point represents the direction in which the gray value gradually changes starting from the starting point. Therefore, choose the adjacent point with the smallest difference as the direction point and arrange it in the sorting direction. The obtained neighborhood distribution sequence reflects the characteristics of the gradual change of gray values around the central pixel point.

[0057] In this embodiment, the absolute value of the difference is used to calculate the difference, and the obtained neighborhood distribution sequence reflects the characteristics of the fundus structure at the pixel point position.

[0058] Since the fundus image contains structures such as the lens, blood vessels, retina, macula, and optic disc, there are obvious differences between various structures, and there is a certain physical connection between the same structures in the image. Therefore, the difference between the gray values of the pixel points in the rotationally oblique image and their surrounding neighborhood pixel points is small. When there is noise in the neighborhood of the central pixel point or the central pixel point is at the edge of structures such as blood vessels, macula, and optic disc edge in the rotationally oblique image, the difference between the gray values of the central pixel point and only some of the neighborhood pixel points is small. In this case, there is a certain difference between the pixel points on both sides of the neighborhood distribution sequence. Therefore, according to the gray value difference between the adjacent pixel points on both sides of the neighborhood distribution sequence, the neighborhood distribution sequence is divided to identify the fundus structure.

[0059] First, according to the gray value difference between the adjacent pixel points on both sides of the neighborhood distribution sequence, calculate the recognition degree of the fundus structure between adjacent pixel points. For the neighborhood distribution sequence of each pixel point, denote the recognition degree of the fundus structure between the i-th and the (i + 1)-th pixel points in the neighborhood distribution sequence as , ;Among them, , respectively represent the gray values of the j-th and k-th pixel points in the neighborhood distribution sequence; i represents the serial number of the previous pixel point among the current two adjacent pixel points in the neighborhood distribution sequence; n represents the number of pixel points in the neighborhood distribution sequence, n = 8; || is the absolute value symbol.

[0060] When the difference between the pixel points on both sides of the neighborhood distribution sequence is larger, that is, the absolute value of the difference between the mean value of the first i pixel points and the mean value of the last n - i pixel points is larger, it indicates that there are obvious differences between the two sides of the neighborhood distribution sequence. This kind of difference is formed by different types of fundus structures. Thus, the recognition degree of the fundus structure between the i-th and (i + 1)-th pixel points in the neighborhood distribution sequence is larger.

[0061] The position with the largest recognition degree of the fundus structure in the neighborhood distribution sequence indicates that there are obvious differences between the pixel points on both sides of this position. The two sides of the sequence are formed by different structures. In order to distinguish different fundus structures, the position with the largest recognition degree of the fundus structure in the neighborhood distribution sequence is used to segment the neighborhood distribution sequence. Specifically, in the neighborhood distribution sequence, the previous and the next pixel points among the pair of adjacent pixel points with the largest recognition degree of the fundus structure are respectively denoted as the previous structure discrimination point and the subsequent structure discrimination point;

[0062] Obtain the mean value of the gray values of all pixel points from the first pixel point in the neighborhood distribution sequence to the previous structure discrimination point, denoted as the first structure feature value; obtain the mean value of the gray values of all pixel points from the subsequent structure discrimination point to the last pixel point in the neighborhood distribution sequence, denoted as the second structure feature value; respectively obtain the differences between the gray value of each pixel point and the first structure feature value and the second structure feature value, denoted as the first difference and the second difference; take the minimum value between the first difference and the second difference as the fundus structure difference degree of each pixel point.

[0063] For example, if the recognition degree of the fundus structure between the q-th and (q + 1)-th pixel points in the neighborhood distribution sequence is the largest, then the mean value of the gray values of the first q pixel points is the first structure feature value, and the mean value of the gray values of the last n - q pixel points is the second structure feature value.

[0064] When the pixel points in the rotational strabismus image are normal pixel points, there is a certain similarity between the pixel points and their local neighborhood pixel points. Since the pixel points may be located at the edge part of the structure in the fundus image, therefore, there are also some dissimilar pixel points between the pixel points in this part and the neighborhood. Thus, the minimum value of the differences between the central pixel point and different structures is taken as the fundus structure difference degree of this central pixel point.

[0065] The block diagram for obtaining the fundus structure difference degree is as Figure 2 shown.

[0066] S3. Based on the difference distribution between the fundus structure dissimilarities of each pixel and all its eight-neighborhood pixels, obtain the number of neighborhood similar structures for each pixel; based on the number of neighborhood similar structures of each pixel, obtain the adaptive smoothing parameter of each pixel during the filtering process.

[0067] For normal pixels, there are pixels in the neighborhood of the pixel that have similar or identical structures to this pixel. The more similar the structures between two pixels are, the smaller the difference between the fundus structure dissimilarities between the pixels. The larger the number of similar pixels in the neighborhood, the more normal the pixel is. Thus, judge the structural characteristics of the pixel according to the difference between the fundus structure dissimilarities between the central pixel and its eight-neighborhood pixels.

[0068] Obtain the difference between the fundus structure dissimilarities of each pixel and its eight-neighborhood pixels as the fundus structure dissimilarity between each pixel and its eight-neighborhood pixels.

[0069] When the difference between the fundus structure dissimilarities between the central pixel and its neighborhood pixels is smaller, it indicates that the structural characteristics between the two pixels are similar, and the two pixels have similar attributes. The more pixels with similar structural characteristics, the more likely the central pixel is a normal pixel, and the denoising intensity for this central pixel should be reduced to ensure the detailed attributes of the pixel.

[0070] Perform Otsu threshold analysis on the fundus structure dissimilarities of all pixels and their eight-neighborhood pixels to obtain the segmentation threshold; specifically, use the fundus structure dissimilarities of all pixels and their eight-neighborhood pixels as the input of the OTSU method, and the output is the segmentation threshold of the fundus structure dissimilarity. Usually, the gray-scale difference between pixels within the same structure is small, making the fundus structure dissimilarity between pixels within the same structure small. Therefore, obtain the fundus structure dissimilarities of all pixels whose fundus structure dissimilarities are less than or equal to the threshold, and form a fundus structure similarity set with them. Among them, the OTSU method is a well-known technology, and the specific calculation process will not be elaborated.

[0071] Furthermore, count the number of fundus structure dissimilarities between each pixel and its eight-neighborhood pixels that are less than the segmentation threshold, and record it as the number of neighborhood similar structures of each pixel. When the number of neighborhood similar structures of a pixel is larger, it indicates that the data point has more similar points locally.

[0072] When the number of neighborhood similar structures of a pixel is larger, it indicates that the pixel is distributed in the same structure of the fundus image, and the pixel has a higher smoothness in the local area. When denoising this pixel, a smaller denoising intensity should be selected to retain more structural information, so that when detecting the patient's rotational strabismus, the detection is more accurate.

[0073] For the NLM filtering algorithm, the smoothing parameter h is used to control the degree of noise suppression and the degree of preservation of image details during the denoising process of the algorithm. If more structural features need to be preserved, the smoothing parameter h needs to be adjusted smaller, so that the pixel points have more detailed attributes during the NLM filtering process; conversely, if the pixel points are noise points, a larger smoothing parameter h should be adopted to enhance the filtering effect.

[0074] Therefore, in this embodiment, the adaptive smoothing parameter of the pixel point is calculated through the number of neighborhood similar structures of the pixel point. Denote the adaptive smoothing parameter of the a-th pixel point in the rotational strabismus image as , , in the formula, represents the preset initial smoothing parameter in the NLM algorithm, and its value in this scheme is 12; A represents the preset adjustment factor, and its value is 9; represents the number of neighborhood similar structures of the a-th pixel point in the rotational strabismus image; represents the sigmoid normalization function.

[0075] When the value of the adaptive smoothing parameter of the pixel point in the rotational strabismus image is larger, it indicates that the number of neighborhood similar structures in its local range is smaller, and the pixel point is more likely to be a noise point. The filtering effect should be increased for this pixel point, so that the pixel value of this pixel point returns to the normal range.

[0076] S4. Filter the rotational strabismus image according to the adaptive smoothing parameter of each pixel point to obtain a rotational strabismus measurement image, detect the circles in the rotational strabismus measurement image to obtain the macula and the optic disc; measure the strabismus degree according to the positions of the macula and the optic disc.

[0077] Take the rotational strabismus image as the input of the NLM algorithm, and use the adaptive smoothing parameter of the pixel point as the smoothing parameter of each pixel point during the filtering process of the NLM algorithm. Among them, the side length of the search window of the algorithm is 21, the side length of the neighborhood window is 3, and the output is the filtered rotational strabismus image, denoted as the rotational strabismus measurement image. Among them, the NLM algorithm is a well-known technology, and the specific calculation process will not be elaborated here.

[0078] For the rotational strabismus measurement image calculated in the above steps, use the Hough transform to fit the circles in the rotational strabismus measurement image. The largest circle is the eyeball, the second largest circle in diameter is the macula, and the smallest circle in diameter is the optic disc. Among them, the Hough transform is a well-known technology, and the specific calculation process will not be elaborated here.

[0079] Finally, the strabismus degree is measured based on the positions of the macula and the optic disc. The calculation of the rotary strabismus degree is a well-known technique, specifically including: in the fundus image of a normal non-rotary strabismus, the position of the macula is at the lower-middle 1 / 3 of the optic disc. To measure the rotary strabismus degree of a patient, the enhanced rotary strabismus measurement image is imported into the professional drawing software Auto CAD, and the center position of the second largest circle is taken as the center position of the macula. The length of the smallest circle is measured by the professional drawing software and trisected. The position where the marking point of the lower-middle 1 / 3 of the smallest circle is located is taken as the position of the lower-middle 1 / 3 point of the optic disc. A measurement straight line is formed by connecting the center of the macula circle. The angle between the measurement straight line and the horizontal line at the lower-middle 1 / 3 of the optic disc in the enhanced rotary strabismus measurement image is the rotary strabismus degree of the patient's eye.

[0080] Doctors can obtain the rotary strabismus degree of patients as an aid through the above steps to achieve the measurement of the rotary strabismus degree of patients.

[0081] In summary, judging rotary strabismus through fundus photos is a non-invasive examination method that does not require direct contact with the eyeball and does not involve any complex equipment. Only by relying on the positions of the macula and the optic disc in the fundus photo can it be preliminarily determined whether there is rotation of the eyeball. This process is painless for patients and is easy to operate, suitable for most people. In this embodiment, through the processing method of analyzing the characteristics of fundus photos, the recognition accuracy of the macula and the optic disc is improved, and more accurate positions of the two are obtained, thereby improving the accuracy of the strabismus degree calculated through the macula and the optic disc.

[0082] From the description of the embodiments in conjunction with the drawings above, those skilled in the art can understand that for the sake of convenience and brevity of description, only the above division of each functional module is used as an example. In actual applications, the above functions can be allocated to different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above.

[0083] The above content is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art within the technical scope disclosed in this application can easily think of changes or substitutions, which should all be covered within the protection scope of this application.

Claims

1. A method for measuring the strabismus degree of patients with rotary strabismus based on fundus photographs, characterized in that, The method includes the following steps: S1, Collect fundus images and perform grayscale processing to obtain rotated strabismus images; S2, According to the gray value difference situation between the eight-neighborhood pixel points of each pixel point in the rotated strabismus image, obtain the neighborhood distribution sequence of each pixel point, and the neighborhood distribution sequence is used to reflect the gradually changing feature of the gray value around the pixel point; According to the gray value difference between the front and back pixel points of all adjacent pixel points in the neighborhood distribution sequence of each pixel point, obtain the fundus structure difference degree of each pixel point, and the fundus structure difference degree is used to reflect the position of the pixel point in the fundus structure, and the fundus structure difference degree is obtained according to the gray value difference between all pixel points before the front structure discrimination point and all pixel points after the back structure discrimination point, the front structure discrimination point is the previous pixel point in a pair of adjacent pixel points with the largest fundus structure recognition degree in the neighborhood distribution sequence, the back structure discrimination point is the latter pixel point in the pair of adjacent pixel points with the largest fundus structure recognition degree, and the fundus structure recognition degree between each pair of adjacent pixel points is obtained according to the gray value difference between all pixel points before and after each pair of adjacent pixel points in the neighborhood distribution sequence of each pixel point; S3, According to the difference distribution between the fundus structure dissimilarity degrees of each pixel point and its all eight-neighborhood pixel points, obtain the number of neighborhood similar structures of each pixel point; According to the number of neighborhood similar structures of each pixel point, obtain the adaptive smoothing parameter of each pixel point during the filtering process; S4, Perform filtering processing on the rotated strabismus image according to the adaptive smoothing parameter of each pixel point to obtain a rotated strabismus measurement image, detect circles in the rotated strabismus measurement image to obtain the macula and the optic disc; Measure the strabismus degree according to the positions of the macula and the optic disc.

2. The method for measuring the strabismus degree of a patient with rotary strabismus based on fundus photographs according to claim 1, wherein, The method for obtaining the neighborhood distribution sequence of each pixel point includes: Denote any pixel point in the rotated strabismus image as the target point; Obtain the pixel point with the smallest gray value among the eight-neighborhood pixel points of the target point, and denote it as the starting point; Starting from the starting point, arrange all eight-neighborhood pixel points in the sorting direction to obtain the neighborhood distribution sequence of the target point.

3. The method for measuring the strabismus degree of a patient with rotary strabismus based on fundus photographs according to claim 2, wherein The method for obtaining the sorting direction includes: Respectively obtain two pixel points adjacent to the starting point among the eight-neighborhood pixel points, and denote them as the two adjacent points of the starting point; Obtain the difference between the gray value of the starting point and the gray value of each of its adjacent points; Take the adjacent point with the smallest difference as the direction point, where if the differences of the two adjacent points are the same, then select any one of the adjacent points as the direction point; If the direction point is in the clockwise direction of the starting point, the sorting direction is the clockwise direction; Otherwise, the sorting direction is the counterclockwise direction.

4. The method for measuring the strabismus degree of a patient with rotary strabismus based on fundus photographs according to claim 1, wherein The method for obtaining the recognition degree of the fundus structure between each pair of adjacent pixel points includes: recording the recognition degree of the fundus structure between the i-th and the (i + 1)-th pixel points in the neighborhood distribution sequence as , ; where respectively represent the gray values of the j-th and the k-th pixel points in the neighborhood distribution sequence; i represents the serial number of the previous pixel point in the current two adjacent pixel points in the neighborhood distribution sequence; n represents the number of pixel points in the neighborhood distribution sequence; || is the absolute value symbol.

5. The method for measuring the strabismus degree of a patient with rotary strabismus based on fundus photographs according to claim 1, wherein Obtaining the fundus structure difference degree of each pixel point according to the gray value difference between all pixel points before the front structure discrimination point and all pixel points after the back structure discrimination point includes: Obtain the average gray value of all pixels from the first pixel in the neighborhood distribution sequence to the pre-structure discrimination point, denoted as the first structure feature value; obtain the average gray value of all pixels from the post-structure discrimination point to the last pixel in the neighborhood distribution sequence, denoted as the second structure feature value; respectively obtain the differences between the gray value of each pixel and the first structure feature value and the second structure feature value, denoted as the first difference and the second difference; take the minimum value between the first difference and the second difference as the fundus structure difference degree of each pixel.

6. The method for measuring the strabismus degree of a patient with rotary strabismus based on fundus photographs according to claim 1, wherein The method for obtaining the number of neighborhood similar structures of each pixel includes: Obtain a segmentation threshold according to the difference situation of the fundus structure difference degrees between all pixels and their respective eight-neighborhood pixels; Count the number of cases where the fundus structure dissimilarity between each pixel and all its eight-neighborhood pixels is less than the segmentation threshold, denoted as the number of neighborhood similar structures of each pixel.

7. The method for measuring the strabismus degree of a patient with rotary strabismus based on fundus photographs according to claim 6, wherein, The method for obtaining the segmentation threshold includes: Obtain the difference in the fundus structure difference degrees between each pixel and its respective eight-neighborhood pixels as the fundus structure dissimilarity between each pixel and its respective eight-neighborhood pixels; Perform Otsu threshold analysis on the fundus structure dissimilarities between all pixels and their all eight-neighborhood pixels to obtain the segmentation threshold.

8. The method for measuring the strabismus degree of a patient with rotary strabismus based on fundus photographs according to claim 1, wherein, Obtaining the adaptive smoothing parameter of each pixel point in the filtering process according to the number of neighborhood similarity structures of each pixel point includes: denoting the adaptive smoothing parameter of the a-th pixel point in the rotationally squinted image as , , where in the formula, represents a preset initial smoothing parameter; represents a preset adjustment factor; represents the number of neighborhood similarity structures of the a-th pixel point in the rotationally squinted image; represents a sigmoid normalization function.

9. The method for measuring the strabismus degree of a patient with rotary strabismus based on fundus photographs according to claim 1, wherein, The method for obtaining the macula and optic disc includes: Substitute the adaptive smoothing parameter of each pixel in the rotation strabismus image into the NLM filtering algorithm to denoise the rotation strabismus image and obtain a rotation strabismus measurement image; Obtain the circles in the rotation strabismus measurement image; take the inner region of the circle with the second largest diameter as the macula; take the inner region of the circle with the smallest diameter as the optic disc.

Images