Automatic whole-eye image stitching method and system based on swept-frequency OCT

Abstract

The present application relates to the field of image processing technology, and in particular to a method and system for automatic whole-eye image stitching based on swept-frequency optical coherence tomography (OCT), including the following steps: collecting anterior segment images and retinal images from different perspectives in combination with manual guidance; preprocessing the anterior segment images from different perspectives to obtain anterior chamber angle information, rigidly registering the anterior segment images based on the anterior chamber angle information, and fusing the registered anterior segment images to obtain anterior segment OCT stitched images; using synchronously acquired laser fundus images to perform horizontal and vertical registration of retinal images from different perspectives, fusing the registered retinal images and performing curvature correction to obtain a retinal OCT stitched image; and stitching the anterior segment OCT stitched image and the retinal OCT stitched image to obtain a whole-eye image. The present application can achieve seamless stitching of the anterior segment image and the retinal image while ensuring the presentation of high-resolution details, thereby obtaining a more complete and structurally clear whole-eye image.

Description

Technical Field

[0001] The present application relates to the field of image processing technology, and in particular to a method and system for automatically stitching whole-eye images based on swept-frequency OCT. Background Art

[0002] Since its introduction into clinical practice, optical coherence tomography (OCT), a non-invasive tissue imaging technique, has been widely used in ophthalmic clinical diagnosis and research due to its high-resolution cross-sectional imaging capabilities. By utilizing a longer wavelength light source and extremely high temporal coherence, swept-frequency OCT not only enhances tissue penetration but also expands the longitudinal imaging range, enabling wider and deeper imaging of the anterior segment and retina. It excels in axial biometry, providing more comprehensive structural and functional information for clinical use.

[0003] Currently, there are two common methods for acquiring whole-eye images based on swept-frequency OCT. One is the single-scan composite method, which simultaneously captures an anterior segment image and a retinal image during a single swept-frequency OCT scan. A full-eye image is then directly generated using an image composite algorithm. The other is the segmented imaging correction method, which first scans the anterior segment and retinal areas independently, then registers and fuses the two images using a curvature or distortion correction algorithm to reconstruct a full-eye image.

[0004] While both methods can obtain full-eye images, the single-scan composite method often compromises resolution and signal-to-noise ratio to balance imaging depth and field of view, resulting in incomplete local structures and unclear details. While the segmented imaging correction method can maintain high resolution within its respective regions, it is prone to registration errors or structural discontinuities at the junction of the anterior segment and the retina, affecting the overall integrity and coherence of the full-eye image. Therefore, how to achieve seamless splicing of the anterior segment image and the retinal image while ensuring high-resolution detail, to obtain a more complete and structurally clear full-eye image, is a current challenge. Summary of the Invention

[0005] This application provides a method and system for automatic whole-eye image stitching based on swept-frequency OCT, which can achieve seamless stitching of anterior segment images and retinal images while ensuring high-resolution detail presentation, thereby obtaining a more complete and structurally clear whole-eye image. This application provides the following technical solutions:

[0006] In a first aspect, the present application provides a method for automatically stitching whole-eye images based on swept-frequency OCT, the method comprising:

[0007] Combined with manual guidance, anterior segment images and retinal images from different perspectives are collected; retinal images from different perspectives include corresponding laser fundus images;

[0008] Preprocessing the anterior segment images from different perspectives to obtain anterior chamber angle information, performing rigid registration on the anterior segment images based on the anterior chamber angle information, and fusing the registered anterior segment images to obtain an anterior segment OCT stitched image;

[0009] The laser fundus images acquired synchronously are used to register the retinal images at different viewing angles in the horizontal and vertical directions. The registered retinal images are then fused and curvature corrected to obtain a retinal OCT mosaic image.

[0010] The anterior segment OCT stitched image and the retinal OCT stitched image are stitched together to obtain a whole-eye image.

[0011] In a specific embodiment, the acquisition of anterior segment images and retinal images from different perspectives in combination with manual guidance includes:

[0012] The anterior segment images at different viewing angles include anterior segment OCT anteroposterior image, anterior segment OCT left image, and anterior segment OCT right image;

[0013] The retinal images at different viewing angles include a retinal OCT orthotopic image and a corresponding laser fundus image, a retinal OCT left side image and a corresponding laser fundus image, and a retinal OCT right side image and a corresponding laser fundus image.

[0014] In a specific embodiment, preprocessing the anterior segment images from different perspectives to obtain anterior chamber angle information, rigidly registering the anterior segment images based on the anterior chamber angle information, and fusing the registered anterior segment images to obtain an anterior segment OCT stitched image includes:

[0015] Refractive correction was performed on the anterior segment OCT anteroposterior images;

[0016] Preprocess the anterior segment image to obtain a mask image of the effective area including the cornea and iris;

[0017] After obtaining the mask image of each anterior segment image, the upper and lower boundary point sets of the anterior chamber angle are accurately located from the mask image. The RANSAC algorithm is applied to the two sets of upper and lower boundary points respectively through random sampling and internal point consistency test to eliminate outliers that deviate significantly from the overall trend.

[0018] Use the remaining candidate points to perform linear fitting on the upper and lower boundaries to obtain two smooth straight lines representing the two sides of the anterior chamber angle. After completing the linear fitting, calculate the intersection of the two boundary lines. , intersection is the vertex coordinate of the anterior chamber angle, and the inclination angle between any boundary line and the horizontal axis is obtained , intersection and tilt angle Together they constitute the registration parameters required for rigid registration.

[0019] In a specific embodiment, the preprocessing of the anterior segment images from different perspectives to obtain anterior chamber angle information, rigidly registering the anterior segment images based on the anterior chamber angle information, and fusing the registered anterior segment images to obtain an anterior segment OCT stitched image further comprises:

[0020] The anterior segment OCT anteroposterior image is used as the reference image, and the anterior segment OCT left image and the anterior segment OCT right image are used as floating images, and rigid registration is performed according to the obtained registration parameters;

[0021] After completing the rigid registration, the floating image is fine-tuned vertically, and the normalized cross-correlation matching is used to measure the degree of match between the template image and the corresponding area of the floating image. The key area that coincides with the floating image is extracted from the reference image as the template image, and all the pixels in the template image are organized into a row vector in column order. , which is the feature vector of the template image; select a vertical sliding window in the floating image, slide row by row in the corresponding overlapping area to extract the candidate area, and also convert it into a feature vector , which is called the feature vector of the detection area; the similarity between regions is measured by calculating the cosine value of the angle between two feature vectors, as shown below:

[0022] ;

[0023] in, and The eigenvectors are and After the entire sliding process is completed, the cosine values of the angles corresponding to all candidate regions are counted, and the position corresponding to the maximum cosine value of the angle is selected as the position where the template image and the floating image are optimally matched. This position is the vertical offset relative to the initial search starting point of the template image. This offset is used as the compensation translation amount to perform a longitudinal translation operation on the floating image.

[0024] The registered anterior segment images were fused using a grayscale weighted method to generate anterior segment OCT stitching images.

[0025] In a specific embodiment, the method of registering retinal images of different viewing angles in the horizontal and vertical directions using the synchronously acquired laser fundus images, fusing the registered retinal images and performing curvature correction to obtain a retinal OCT stitched image includes:

[0026] The retinal OCT orthotopic image is used as the reference image, and the retinal OCT left image and the retinal OCT right image are used as floating images. For each of the retinal OCT orthotopic image and the retinal OCT left image or the retinal OCT right image, it corresponds to a laser fundus image. The two laser fundus images are used as the registration input, and the scale-invariant feature transformation algorithm is used to extract the feature points of the laser fundus image. The extracted feature points are matched using a fast nearest neighbor search package to find similar feature point pairs.

[0027] The matching points are screened using the RANSAC algorithm. Based on the screened valid feature point pairs, the mean of their horizontal distances is calculated to obtain the lateral overlap length of the laser fundus image. Calculate the horizontal overlap length of the retinal image based on the horizontal overlap length of the laser fundus image as follows:

[0028] ;

[0029] in, is the image width of the laser fundus image in the registration center, is the image width of the corresponding retinal OCT image, according to the horizontal overlap length of the retinal image , determining the horizontal alignment position of the reference image and the floating image, and adjusting the position of the floating image in the horizontal direction;

[0030] The column pixel position corresponding to the reference image and the floating image at the splicing point is determined based on the calculated horizontal overlap length. Based on this column of pixels, the vertical displacement difference between the reference image and the floating image is calculated to obtain the vertical translation parameter. The vertical translation transformation operation is performed on the floating image according to the vertical translation parameter to achieve vertical position alignment.

[0031] In a specific embodiment, the method of registering retinal images of different viewing angles in the horizontal and vertical directions using the synchronously acquired laser fundus images, fusing the registered retinal images and performing curvature correction to obtain a retinal OCT stitched image further includes:

[0032] The registered retinal images were fused using a grayscale weighted method to generate a retinal OCT mosaic image.

[0033] According to the retinal OCT imaging characteristics and optical structure model, the curvature of the retinal OCT mosaic image is corrected to generate a retinal OCT mosaic image close to the actual fundus curvature;

[0034] Curvature correction maps the stitched image into a sector image structure from the posterior end of the vitreous to the fundus, with the center of the sector being the center of the circle. The scanning axis point at the pupil is used to calculate the opening angle corresponding to the complete stitched image based on the stitching width of the image and the optical parameters of OCT imaging. as follows:

[0035] ;

[0036] in, is the scanning field angle of the retinal OCT image, is the width of the retinal OCT stitching image, is the image width of the retinal OCT image;

[0037] For any point in the fan ring structure , from the center of the circle Physical distance The calculation formula is as follows based on the optical path and the refractive index of the medium:

[0038] ;

[0039] in, and They are The optical path thickness of the lens and vitreous body at the point position, and are the refractive indices of the lens and vitreous humor, respectively;

[0040] After the calculation is completed, each pixel in the retinal OCT mosaic image is combined with its relative opening angle The determined scanning angle and the corresponding physical distance , transform from the Cartesian coordinate system to the polar coordinate system with the scanning axis point as the pole to obtain its true spatial position distribution, and then through the inverse transformation from polar coordinates to Cartesian coordinates, map the retinal OCT stitching image back to the plane image to complete the correction of curvature distortion.

[0041] In a specific embodiment, stitching the anterior segment OCT stitched image and the retinal OCT stitched image to obtain a whole-eye image includes:

[0042] Adjust the size of the retinal OCT mosaic image to keep it consistent with the physical resolution of the anterior segment OCT mosaic image;

[0043] When stitching images, find the intersection of the visual axis and the line connecting the two anterior chamber angles in the anterior segment OCT stitching image, and determine the relative positions of the anterior segment OCT stitching image and the retinal OCT stitching image based on the coordinates of the intersection;

[0044] The anterior segment OCT stitching image and the retinal OCT stitching image are moved and stitched according to their relative positions to obtain a whole-eye stitching image.

[0045] In a second aspect, the present application provides a system for automatically stitching whole-eye images based on swept-frequency OCT, which adopts the following technical solutions:

[0046] A system for automatically stitching whole-eye images based on swept-frequency OCT, comprising:

[0047] An image acquisition module, used to acquire anterior segment images and retinal images from different perspectives in combination with manual guidance;

[0048] an anterior segment image stitching module, configured to pre-process the anterior segment images of different viewing angles to obtain anterior chamber angle information, perform rigid registration on the anterior segment images based on the anterior chamber angle information, and fuse the registered anterior segment images to obtain an anterior segment OCT stitching image;

[0049] The retinal image stitching module is used to register retinal images of different viewing angles in the horizontal and vertical directions using the synchronously acquired laser fundus images, fuse the registered retinal images, and perform curvature correction to obtain a retinal OCT stitching image;

[0050] The whole-eye image stitching module is used to stitch the anterior segment OCT stitching image and the retinal OCT stitching image to obtain a whole-eye image.

[0051] In a third aspect, the present application provides an electronic device comprising a processor and a memory; the memory stores a program, which is loaded and executed by the processor to implement a method for automatic whole-eye image stitching based on swept-frequency OCT as described in the first aspect.

[0052] In a fourth aspect, the present application provides a computer-readable storage medium, wherein the storage medium stores a program, and when the program is executed by a processor, it is used to implement a method for automatic stitching of whole-eye images based on swept-frequency OCT as described in the first aspect.

[0053] In summary, the beneficial effects of this application include at least:

[0054] (1) On the one hand, the multi-view acquisition strategy combined with manual guidance ensures at least 20% overlap between each pair of adjacent images, providing sufficient feature information for subsequent registration; on the other hand, through the anterior chamber angle feature extraction of the anterior segment image, RANSAC-enhanced rigid registration and grayscale weighted fusion algorithm, not only the high-resolution details of tiny tissue structures such as the cornea and iris are retained in the region, but also the grayscale mutation and distortion artifacts at the splicing edge are eliminated, so that the spliced images of the anterior segment and retina achieve a natural and smooth transition at the detail level, thus solving the difficult problem of the compromise between resolution and signal-to-noise ratio in a single scan.

[0055] (2) To address the problem of unclear features and high noise in retinal OCT images, synchronously acquired laser fundus images are introduced to assist in completing highly robust horizontal and vertical registration; then, through the curvature correction method, the spatial form of the scanning sector ring on the real sphere is restored based on the optical path and refractive index of the lens and vitreous body; finally, the scanning axis point at the pupil is used as the common positioning reference to unify the physical scale of the anterior segment and retinal images and seamlessly splice them, so that the final full-eye image is not only visually coherent and natural, but also geometrically fits the inner and outer curvatures of the real eyeball, providing a highly consistent and reliable data basis for subsequent three-dimensional reconstruction and quantitative analysis.

[0056] By combining manual guidance to first collect anterior segment images and retinal images from different perspectives, the incomplete coverage and registration difficulties caused by limited field of view are eliminated from the source; then the anterior segment images from different perspectives are preprocessed to obtain the anterior chamber angle information, and the anterior segment images are rigidly registered based on the anterior chamber angle information, and the registered anterior segment images are fused to obtain the anterior segment OCT stitching image; this not only retains high-resolution tissue details, but also achieves a natural transition between regions, solving the problem of unclear details caused by the compromise between single scan resolution and signal-to-noise ratio; then the laser fundus images acquired synchronously are used to align the anterior segment images from different perspectives. The retinal images are aligned horizontally and vertically, and the aligned retinal images are fused and curvature corrected to obtain a retinal OCT stitched image. Laser fundus images are introduced to assist in completing highly robust horizontal and vertical registration, and the true fundus curvature is restored through spherical curvature correction, which effectively compensates for the structural discontinuity and distortion at the stitching junction of the traditional segmented correction method; finally, on the basis of unified physical scale and scanning axis positioning, the anterior segment and retinal stitched images are seamlessly synthesized, achieving a balance between high resolution and overall coherence, thereby obtaining a whole-eye image with more complete structure and clearer details.

[0057] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application and to implement it in accordance with the contents of the specification, the following is a detailed description of the preferred embodiments of the present application in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 It is a schematic diagram of the overall process of the automatic whole-eye image stitching method based on swept-frequency OCT in an embodiment of the present application.

[0059] Figure 2 This is a schematic diagram of a use case of anterior segment images and retinal images collected in an embodiment of the present application.

[0060] Figure 3 It is a flowchart of step S102 in the embodiment of the present application.

[0061] Figure 4 This is a schematic diagram of a use case for determining the upper and lower boundary points of the anterior chamber angle in an embodiment of the present application.

[0062] Figure 5 It is a flowchart of step S103 in the embodiment of the present application.

[0063] Figure 6 This is a schematic diagram of a use case of curvature correction in an embodiment of the present application to form a fan-shaped ring image from the posterior end of the vitreous body to the fundus.

[0064] Figure 7 This is a schematic diagram of use cases of images after anterior segment stitching, retinal stitching, and whole eye stitching in the embodiments of the present application.

[0065] Figure 8 This is a structural block diagram of the full-eye image automatic stitching system based on swept-frequency OCT in an embodiment of the present application.

[0066] Figure 9 This is a block diagram of an electronic device for automatically stitching whole-eye images based on swept-frequency OCT in an embodiment of the present application. DETAILED DESCRIPTION

[0067] The following embodiments are used to illustrate the present invention, but are not intended to limit the scope of the present invention.

[0068] Optionally, the present application uses the automatic stitching method of full-eye images based on swept-frequency OCT provided in each embodiment as an example for description when used in an electronic device, where the electronic device is a terminal or a server. The terminal can be a mobile phone, a computer, a tablet computer, etc. This embodiment does not limit the type of electronic device.

[0069] Reference Figure 1 , is a flow chart of a method for automatic whole-eye image stitching based on swept-frequency OCT provided in one embodiment of the present application, which method includes at least the following steps:

[0070] Step S101: Acquire anterior segment images and retinal images from different perspectives in combination with manual guidance.

[0071] In step S101, combined with manual guidance, images of the anterior segment and retina of the subject are sequentially captured from different viewing angles. In practice, whole-eye image capture is divided into two parts: anterior segment image capture and retinal image capture. Due to the limited field of view of a single OCT image, the maximum external field of view for anterior segment OCT is typically 39°, while the maximum external field of view for retinal OCT is typically 90°. Considering that the overlapping area for subsequent image stitching requires at least 20% overlap, at least three anterior segment OCT images and three retinal OCT images must be captured from three viewing angles, namely, anteroposterior, leftward, and rightward, to generate a whole-eye image covering the entire meridian plane.

[0072] Specifically, during the acquisition of the anterior segment image, the anterior segment OCT anteroposterior image is first acquired under the maximum imaging field of view, and then, under the same mode, the position of the fixation lamp is adjusted to guide the subject to look horizontally to the left and horizontally to the right, respectively, to obtain the anterior segment OCT left image and the anterior segment OCT right image. Similarly, during the acquisition of the retinal image, the retinal OCT anteroposterior image and the corresponding laser fundus image are first synchronously acquired under the maximum imaging field of view, and then, under the same mode, the position of the fixation lamp is adjusted to guide the subject to look horizontally to the left and horizontally to the right, respectively, to obtain the retinal OCT left image and the corresponding laser fundus image, and the retinal OCT right image and the corresponding laser fundus image. Figure 2 From left to right, from top to bottom are a set of images of the anterior segment, anterior segment left side, anterior segment right side, retina, retina right side, retina left side, and retina right side collected by whole-eye stitching. Among them, the anterior segment only needs OCT images, and the retina needs OCT images and corresponding laser fundus images.

[0073] It should be noted that the purpose of acquiring three images, one from the frontal, left, and right viewpoints, is to ensure sufficient overlap between the two adjacent viewpoints, providing stable registration feature points. A minimum overlap of 20% ensures that the stitching algorithm can find sufficient common structure (such as the corneal edge, lens outline, or retinal vascular branches) between the images, achieving a seamless connection and maintaining overall structural continuity. However, using only two images often fails to simultaneously meet the full field of view coverage and overlap requirements. Therefore, at least three images are required to cover the entire meridian plane and ensure a stable stitching foundation between each pair of adjacent images.

[0074] Step S102 : pre-processing the anterior segment images of different viewing angles to obtain anterior chamber angle information, rigidly registering the anterior segment images based on the anterior chamber angle information, and fusing the registered anterior segment images to obtain an anterior segment OCT spliced image.

[0075] Reference Figure 3, is a flow chart of step S102 in an embodiment of the present application. The method includes at least the following steps:

[0076] Step S1021: performing refractive correction on the anterior segment OCT anteroposterior image in the anterior segment image.

[0077] Specifically, in order to eliminate the impact of refractive error on subsequent measurements and stitching, refractive correction is performed on the orthotopic image to ensure more accurate measurement of subsequent physiological parameters. In the anterior segment stitching process, the main purpose of refractive correction is to eliminate the geometric distortion caused by differences in the refractive index of the eyeball, thereby ensuring the accuracy of the measurement results. These physiological parameters are usually only extracted from the orthotopic image where the central visual axis is located, and the lateral image is more used to provide structural overlapping information to complete the stitching alignment. Therefore, only performing refractive correction on the orthotopic image can not only meet the measurement accuracy requirements, but also avoid unnecessary optical correction of the lateral image, thereby simplifying the process.

[0078] Step S1022: pre-process the front segment image to obtain a mask image of the effective area including the cornea and iris.

[0079] Specifically, the three collected anterior segment images were subjected to contrast stretching and gamma enhancement to increase the contrast of the anterior segment images. The enhanced images were then binarized using the maximum inter-class variance method to obtain mask images of the effective area including the cornea and iris.

[0080] Step S1023: extract the upper and lower boundary point sets of the anterior chamber angle based on the mask image of each anterior segment image, and obtain the registration parameters by linear fitting.

[0081] In step S1023, after obtaining the mask image of each anterior segment image, the upper and lower boundary point sets of the anterior chamber angle are accurately located from the mask image. The anterior chamber angle is the angle between the cornea and the iris, and the alignment parameters required for subsequent rigid alignment are obtained by linear fitting.

[0082] In the implementation, the anterior chamber angle boundary points of the mask image obtained by preprocessing are searched in columns. , specifically, refer to Figure 4 , which is a schematic diagram of a use case for determining the upper and lower boundary points of the anterior chamber angle in an embodiment of the present application, searching for the first consecutive The first pixel is 255, followed by a continuous The coordinates of the point with pixel 0 are , then the upper boundary point of the anterior chamber angle in this column ; Then search for the first consecutive The coordinates of the point where n pixels are 255 and then n consecutive pixels are 0 are , then the lower boundary point of the anterior chamber angle in this column ,and That is, for each mask image in the previous section, perform pixel scanning column by column from left to right. First, in each column, search from the top downward for the first position that satisfies the requirement of n consecutive foreground pixels followed by n background pixels. Record the upper boundary point of the column as the position offset n–1 rows downward from the start of the foreground. Then, search from the bottom upward in the same column for a starting position that meets the same conditions. Record the lower boundary point as the position offset n–1 rows upward from this position. In this way, a set of upper and lower boundary candidate points distributed in the column direction can be obtained.

[0083] In addition, since there may be noise or local breaks in the mask image, directly using all candidate points may lead to unstable fitting results. To this end, the RANSAC algorithm is first applied to the two sets of points at the upper and lower boundaries respectively through random sampling and internal point consistency test to effectively eliminate outliers that deviate greatly from the overall trend. Subsequently, the remaining candidate points are used to perform linear fitting on the upper and lower boundaries respectively to obtain two smooth straight lines representing the two sides of the anterior chamber angle. After completing the linear fitting, the intersection of the two boundary lines is calculated. , the intersection Reflects the vertex coordinates of the anterior chamber angle and obtains the inclination angle between any boundary line and the horizontal axis The above intersection point and tilt angle together constitute the registration parameters required for rigid registration: the intersection point is used as the rotation center and reference positioning, and the tilt angle is used for subsequent image rotation correction.

[0084] Step S1024 : Use the anterior segment OCT anteroposterior image as a reference image, the anterior segment OCT left image and the anterior segment OCT right image as floating images, and perform rigid registration according to the obtained registration parameters.

[0085] In step S1024, the anterior segment OCT anteroposterior image among the three anterior segment images is used as the reference image, the anterior segment OCT left image and the anterior segment OCT right image are used as floating images, and the floating images are rigidly registered based on the registration parameters obtained in step S1023 to achieve consistent alignment of the three images in spatial position and angle.

[0086] Specifically, the registration parameters include the coordinates of the vertex of the anterior chamber angle in each anterior segment image and the inclination angle with respect to the horizontal axis During the registration process, the anterior chamber angle vertex in the floating image is first compared with the reference image. The coordinates of the anterior chamber angle are calculated, and their horizontal and vertical displacement differences are calculated. Based on this, the floating image is translated so that the anterior chamber angle vertex is aligned with the reference image in spatial position. Then, with the aligned anterior chamber angle vertex as the rotation center, the inclination angle of the anterior chamber angle boundary line in the floating image and the reference image is calculated. The difference is used as the rotation angle to adjust the posture direction of the floating image to achieve structural angle alignment.

[0087] Through the above translation and rotation operations, the floating image is rigidly transformed so that it is precisely aligned with the reference image in terms of structural features, ensuring that the three images have a unified geometric reference frame, providing a standardized basis for subsequent image fusion, analysis and feature extraction.

[0088] Furthermore, after completing the aforementioned rigid transformation, the floating image is preferably fine-tuned longitudinally to further improve the alignment accuracy of local structures between the images. Because actual images may exhibit slight vertical offsets during capture, rigid registration based solely on the anterior chamber angle may not completely eliminate all displacement errors. Therefore, a correlation-based template matching method is introduced to match the overlapping regions of the reference image and the floating image to obtain accurate longitudinal translation compensation.

[0089] Specifically, the normalized cross-correlation matching is used to measure the matching degree between the template image and the corresponding area of the floating image. First, the key area that coincides with the floating image is extracted from the reference image as the template image, and all the pixels in the template image are organized into a row vector in column order. , which is the feature vector of the template image. Then, a vertical sliding window is selected in the floating image, and the candidate area is extracted row by row in the corresponding overlapping area, and is also converted into a feature vector , which is called the feature vector of the detection area. The similarity between regions is measured by calculating the cosine value of the angle between two feature vectors, as shown below:

[0090] ;

[0091] in, and The eigenvectors are and The Euclidean norm of the angles is calculated. After the entire sliding process is complete, the cosine values of the angles corresponding to all candidate regions are counted, and the position corresponding to the maximum cosine value is selected as the location where the template image and the floating image are optimally matched. The vertical offset of this position relative to the initial search starting point of the template image is the number of pixels by which the floating image should be adjusted vertically. This offset is used as the compensation translation, and the floating image is vertically translated, ultimately achieving precise alignment of the overlapping regions, further improving the registration quality of the entire image in terms of local details.

[0092] Step S1025 : Using a grayscale weighted method to fuse the registered anterior segment images to generate an anterior segment OCT spliced image.

[0093] In step S1025, the reference image and the floating image, after rigid registration and fine-tuning, are fused using a grayscale-weighted method to generate a complete, continuous, and naturally transitioned anterior segment OCT spliced image. This fusion operation is primarily applied to the overlapping regions between the anteroposterior image and its left and right floating images, aiming to eliminate artifacts such as grayscale abrupt changes and edge breaks that may exist at the image splicing site.

[0094] Specifically, the overlap area on one side of the orthogonal image Overlapping area with its corresponding lateral image , the overlapping area of the fused images It can be expressed as:

[0095] ;

[0096] in, represents the corresponding overlapping area in the orthogonal image, Indicates the corresponding overlapping area in the floating image (left or right), and Represents images respectively and The weight coefficient of In the overlapping area on the other side, the fusion process is also performed according to the above weighted method.

[0097] Step S103: Use the synchronously acquired laser fundus images to perform horizontal and vertical registration on the retinal images of different viewing angles, fuse and perform curvature correction on the registered retinal images to obtain a retinal OCT spliced image.

[0098] Reference Figure 5 , is a flow chart of step S103 in an embodiment of the present application. The method includes at least the following steps:

[0099] Step S1031: Using the retinal OCT normal image as a reference image, the retinal OCT left image and the retinal OCT right image as floating images, and using the synchronously acquired laser fundus image, perform horizontal and vertical registration on the retinal image.

[0100] In step S1031, the purpose is to use the simultaneously acquired laser fundus images to assist in the horizontal and vertical registration of the retinal images. The retinal OCT orthotopic image is used as the reference image, and the left and right retinal OCT images are used as floating images. Direct retinal image registration is difficult due to the high noise and unclear features of retinal images from different viewing angles. Therefore, the retinal OCT images are indirectly registered using the simultaneously acquired laser fundus images.

[0101] Specifically, for any one of the retinal OCT orthotopic images and the retinal OCT left image or right image, it corresponds to a laser fundus image acquired synchronously. Using these two laser fundus images as registration inputs, the scale-invariant feature transformation algorithm is used to extract the feature points of the laser fundus image, and then the extracted feature points are matched through the fast nearest neighbor search package to find similar feature point pairs. Then, the RANSAC algorithm is used to filter the matching points and remove the wrong matching points to ensure the reliability of the matching point pairs. Based on the valid feature point pairs after screening, the mean of their horizontal distances is calculated to obtain the lateral overlap length of the laser fundus image. Calculate the horizontal overlap length of the retinal image based on the horizontal overlap length of the laser fundus image as follows:

[0102] ;

[0103] in, is the image width of the laser fundus image in the registration center (the width of the left laser fundus image or the right laser fundus image), is the image width of the corresponding retinal OCT image (for example, the width of the left or right retinal OCT image). , determine the horizontal alignment position of the reference image and the floating image, and adjust the position of the floating image in the horizontal direction to ensure that it is accurately aligned with the reference image in the horizontal direction.

[0104] After completing horizontal registration, vertical registration is performed. Specifically, the corresponding column pixel position at the splicing point between the reference image and the floating image is determined based on the horizontal overlap length calculated above. Based on this column pixel position, a cross-correlation-based template matching method is used to calculate the vertical displacement difference between the reference image and the floating image, obtaining a vertical translation parameter. Based on this vertical translation parameter, a vertical translation transformation operation is performed on the floating image to achieve vertical alignment.

[0105] It should be noted that the cross-correlation-based template matching method has the same principle as step S1024 and will not be described in detail here.

[0106] It should be noted that the above-mentioned horizontal and vertical registration operations are performed once for the left image and once for the right image respectively, and finally the three retinal images are accurately registered in a unified coordinate system.

[0107] Step S1032: fuse the registered retinal images using a grayscale weighted method to generate a retinal OCT mosaic image.

[0108] It should be noted that the principle here is the same as that of step S1025 and will not be described in detail here.

[0109] Step S1033: Perform curvature correction on the retinal OCT stitched image according to the retinal OCT imaging characteristics and the optical structure model to generate a retinal OCT stitched image that is close to the actual fundus curvature.

[0110] In step S1033, during the retinal OCT imaging process, the scanning light beam scans and images the fundus in a fan-shaped manner with the scanning axis point at the pupil as the center. Therefore, each frame of OCT image can be approximately regarded as an equal phase surface perpendicular to the scanning light. In three-dimensional space, these surfaces are actually spherical slices formed around the scanning axis point. This imaging method causes the retinal structure in the image to present a certain curved shape. Therefore, if multiple retinal images are spliced together, if they are directly analyzed or three-dimensionally reconstructed, the image that has not been curvature corrected will not accurately reflect the actual spatial structure of the fundus. Therefore, it is necessary to perform curvature correction on the spliced image according to the OCT imaging optical model so that its geometric shape fits the real fundus sphere, thereby improving spatial consistency and analysis accuracy.

[0111] Specifically, if Figure 6 As shown in Figure 2, the essence of curvature correction is to map the stitched image into a sector image structure from the posterior end of the vitreous to the fundus. The center of the circle where the sector is located is The scanning axis point at the pupil is the concentric circle of the circle, which represents the equal optical path surface. According to the image stitching width and the optical parameters of OCT imaging, the opening angle corresponding to the complete stitching image is calculated. as follows:

[0112] ;

[0113] in, is the scanning field angle of the retinal OCT image, is the width of the retinal OCT stitching image, is the image width of the retinal OCT image. Used to determine how far the entire stitched image spans within the field of view.

[0114] Next, for any point in the fan ring structure , from the center of the circle Physical distance The calculation formula is as follows based on the optical path and the refractive index of the medium:

[0115] ;

[0116] in, and They are The optical path thickness of the lens and vitreous body at the point position, and are the refractive indices of the lens and vitreous humor, respectively.

[0117] After the calculation is completed, each pixel in the retinal OCT mosaic image is combined with its relative opening angle The specific scanning angle and the corresponding physical distance determined , transforming from the Cartesian coordinate system to the polar coordinate system with the scanning axis as the pole to obtain its true spatial position distribution. Subsequently, through the inverse transformation from polar coordinates to Cartesian coordinates, the retinal OCT mosaic image is mapped back to a planar image, thus completing the correction of curvature distortion.

[0118] In order to determine the specific scanning angle corresponding to each pixel in the spliced image, the opening angle is first calculated based on , establish a correspondence between the width range of the stitched image and the field of view angle range. That is, the leftmost pixel in the stitched image corresponds to the starting angle of the opening angle range, the rightmost pixel corresponds to the ending angle of the opening angle range, and the remaining pixels are mapped to the opening angle range in a linear proportion according to their relative horizontal positions in the stitched image. In this way, a specific scanning angle can be assigned to each pixel in the stitched image, so that in the subsequent conversion process between the Cartesian coordinate system and the polar coordinate system, the corresponding physical distance Achieve accurate spatial position reconstruction and curvature correction.

[0119] Step S104: stitching the anterior segment OCT stitching image and the retinal OCT stitching image to obtain a full-eye image.

[0120] In step S104, since the anterior segment and retinal images come from different sources, they usually have different physical resolutions. In order to ensure that the stitched image can be presented at the same scale, it is first necessary to adjust the size of the retinal OCT stitched image to keep it consistent with the physical resolution of the anterior segment OCT stitched image. Then, when stitching the images, it is first necessary to determine the relative positions of the anterior segment OCT stitched image and the retinal OCT stitched image. The positioning relationship between the two is calculated based on the scanning axis point at the pupil. Specifically, find the intersection of the visual axis and the line connecting the anterior chamber angles on both sides in the anterior segment OCT stitched image. This intersection is the same as the center of the circle where the fan ring is located after curvature correction in the retinal OCT stitched image. The relative positions of the anterior segment OCT stitched image and the retinal OCT stitched image are determined based on the coordinates of the intersection. According to the relative positions of the anterior segment OCT stitched image and the retinal OCT stitched image obtained above, the two are moved and stitched to obtain a whole-eye stitched image, such as Figure 7As shown, from top to bottom and from left to right are the anterior segment OCT stitching image, the retinal OCT stitching image and the stitched whole-eye image.

[0121] In summary, this application provides a method for automatic whole-eye image stitching based on swept-frequency OCT, aiming to solve the problem in the prior art of how to achieve seamless stitching of the anterior segment image and the retinal image while ensuring high-resolution detail presentation. Although the existing swept-frequency OCT technology can provide high-resolution anterior segment images and retinal images, due to its field of view limitations and problems with the registration accuracy between images, the stitched image cannot achieve a high-precision whole-eye image effect, and often has defects such as unnatural stitching and loss of details.

[0122] In response to the shortcomings of the existing technology, this application first collects anterior segment images and retinal images from different perspectives by combining manual guidance, thereby eliminating the incomplete coverage and registration difficulties caused by limited field of view from the source; then pre-processes the anterior segment images from different perspectives to obtain anterior chamber angle information, and rigidly registers the anterior segment images based on the anterior chamber angle information, and fuses the registered anterior segment images to obtain anterior segment OCT stitching image; it not only retains high-resolution tissue details, but also achieves a natural transition between regions, and solves the problem of unclear details caused by the compromise between single scan resolution and signal-to-noise ratio; then uses the laser fundus image acquired synchronously For example, retinal images from different perspectives are registered horizontally and vertically, and the registered retinal images are fused and curvature corrected to obtain a retinal OCT stitched image. Laser fundus images are introduced to assist in completing highly robust horizontal and vertical registration, and the true fundus curvature is restored through spherical curvature correction, effectively compensating for the structural discontinuity and distortion at the stitching junction of the traditional segmented correction method; finally, the anterior segment and retinal stitched images are seamlessly synthesized on the basis of unified physical scale and scanning axis positioning, achieving a balance between high resolution and overall coherence, thereby obtaining a whole-eye image with a more complete structure and clearer details.

[0123] Figure 8 This is a block diagram of a system for automatically stitching whole-eye images based on swept-frequency OCT, provided in one embodiment of the present application. The system includes at least the following modules:

[0124] An image acquisition module, used to acquire anterior segment images and retinal images from different perspectives in combination with manual guidance;

[0125] Anterior segment image stitching module, used to pre-process anterior segment images from different perspectives to obtain anterior chamber angle information, perform rigid registration on the anterior segment images based on the anterior chamber angle information, and fuse the registered anterior segment images to obtain anterior segment OCT stitching image;

[0126] The retinal image stitching module is used to register retinal images of different viewing angles in the horizontal and vertical directions using the synchronously acquired laser fundus images, fuse the registered retinal images, and perform curvature correction to obtain a retinal OCT stitching image;

[0127] The whole-eye image stitching module is used to stitch the anterior segment OCT stitching image and the retinal OCT stitching image to obtain a whole-eye image.

[0128] For relevant details, please refer to the above method embodiment.

[0129] Figure 9 4 is a block diagram of an electronic device provided in one embodiment of the present application. The device includes at least a processor 401 and a memory 402.

[0130] Processor 401 may include one or more processing cores, such as a quad-core processor or an octa-core processor. Processor 401 may be implemented in hardware using at least one of the following: a DSP (Digital Signal Processing), an FPGA (Field-Programmable Gate Array), or a PLA (Programmable Logic Array). Processor 401 may also include a main processor and a coprocessor. The main processor is used to process data in the awake state, also known as a CPU (Central Processing Unit); the coprocessor is a low-power processor used to process data in the standby state. In some embodiments, processor 401 may be integrated with a GPU (Graphics Processing Unit), which is responsible for rendering and drawing content displayed on the display screen. In some embodiments, processor 401 may also include an AI (Artificial Intelligence) processor, which is used to handle computational operations related to machine learning.

[0131] The memory 402 may include one or more computer-readable storage media, which may be non-transitory. The memory 402 may also include a high-speed random access memory and a non-volatile memory, such as one or more disk storage devices and flash memory storage devices. In some embodiments, the non-transitory computer-readable storage medium in the memory 402 is used to store at least one instruction, which is executed by the processor 401 to implement the full-eye image automatic stitching method based on swept-frequency OCT provided in the method embodiment of the present application.

[0132] In some embodiments, the electronic device may optionally include a peripheral device interface and at least one peripheral device. The processor 401, memory 402, and peripheral device interface may be connected via a bus or signal lines. Each peripheral device may be connected to the peripheral device interface via a bus, signal lines, or circuit boards. Illustratively, the peripheral devices include, but are not limited to, radio frequency circuitry, a touchscreen display, audio circuitry, and a power supply.

[0133] Of course, the electronic device may also include fewer or more components, which is not limited in this embodiment.

[0134] Optionally, the present application also provides a computer-readable storage medium, in which a program is stored. The program is loaded and executed by a processor to implement the automatic whole-eye image stitching method based on swept-frequency OCT of the above-mentioned method embodiment.

[0135] Optionally, the present application also provides a computer product, which includes a computer-readable storage medium, in which a program is stored. The program is loaded and executed by a processor to implement the automatic whole-eye image stitching method based on swept-frequency OCT of the above-mentioned method embodiment.

[0136] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0137] The above embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art could make various modifications and improvements without departing from the spirit of the present application, all of which fall within the scope of protection of the present application. Therefore, the scope of protection of the present patent application shall be determined by the appended claims.

Claims

1. A method for automatic whole-eye image stitching based on swept-frequency OCT, characterized in that: The method comprises: Anterior segment images and retinal images at different viewing angles are collected in combination with manual guidance; the retinal images at different viewing angles include corresponding laser fundus images; the anterior segment images at different viewing angles include anterior segment OCT orthotopic images, anterior segment OCT left images, and anterior segment OCT right images; the retinal images at different viewing angles include retinal OCT orthotopic images and corresponding laser fundus images, retinal OCT left images and corresponding laser fundus images, and retinal OCT right images and corresponding laser fundus images; Preprocessing the anterior segment images of different viewing angles to obtain anterior chamber angle information, rigidly registering the anterior segment images based on the anterior chamber angle information, and fusing the registered anterior segment images to obtain an anterior segment OCT mosaic image, including: performing refractive correction on the anterior segment OCT anteroposterior image in the anterior segment image; preprocessing the anterior segment image to obtain a mask image of a valid area including the cornea and the iris; extracting the upper and lower boundary point sets of the anterior chamber angle based on the mask image of each anterior segment image, and obtaining registration parameters by linear fitting; using the anterior segment OCT anteroposterior image as a reference image, the anterior segment OCT left image and the anterior segment OCT right image as floating images, and performing rigid registration according to the obtained registration parameters; and fusing the registered anterior segment images using a grayscale weighted method to generate an anterior segment OCT mosaic image; Synchronously acquired laser fundus images are used to perform horizontal and vertical registration of retinal images from different viewing angles, and the registered retinal images are fused and curvature corrected to obtain a retinal OCT mosaic image, including: using the retinal OCT orthotopic image as a reference image, the retinal OCT left image and the retinal OCT right image as floating images, and using the synchronously acquired laser fundus images to perform horizontal and vertical registration of the retinal images; using a grayscale weighted method to fuse the registered retinal images to generate a retinal OCT mosaic image; and performing curvature correction on the retinal OCT mosaic image based on retinal OCT imaging characteristics and an optical structure model to generate a retinal OCT mosaic image that is close to the actual fundus curvature. The anterior segment OCT stitched image and the retinal OCT stitched image are stitched together to obtain a whole-eye image.

2. The method for automatic whole-eye image stitching based on swept-frequency OCT according to claim 1, characterized in that: The mask image based on each anterior segment image extracts the upper and lower boundary point sets of the anterior chamber angle, and obtains the registration parameters by linear fitting, including: After obtaining the mask image of each anterior segment image, the upper and lower boundary point sets of the anterior chamber angle are accurately located from the mask image. The RANSAC algorithm is applied to the two sets of upper and lower boundary points respectively through random sampling and internal point consistency test to eliminate outliers that deviate significantly from the overall trend. Use the remaining candidate points to perform linear fitting on the upper and lower boundaries to obtain two smooth straight lines representing the two sides of the anterior chamber angle. After completing the linear fitting, calculate the intersection of the two boundary lines. , intersection is the vertex coordinate of the anterior chamber angle, and the inclination angle between any boundary line and the horizontal axis is obtained , intersection and tilt angle Together they constitute the registration parameters required for rigid registration.

3. The method for automatic whole-eye image stitching based on swept-frequency OCT according to claim 2, characterized in that: The preprocessing of the anterior segment images at different viewing angles to obtain anterior chamber angle information, rigidly registering the anterior segment images based on the anterior chamber angle information, and fusing the registered anterior segment images to obtain an anterior segment OCT stitched image further comprises: After completing the rigid registration, the floating image is fine-tuned vertically, and the normalized cross-correlation matching is used to measure the degree of match between the template image and the corresponding area of the floating image. The key area that coincides with the floating image is extracted from the reference image as the template image, and all the pixels in the template image are organized into a row vector in column order. , which is the feature vector of the template image; select a vertical sliding window in the floating image, slide row by row in the corresponding overlapping area to extract the candidate area, and also convert it into a feature vector , which is the feature vector of the detection area; the similarity between regions is measured by calculating the cosine value of the angle between two feature vectors, as shown below: ; in, and The eigenvectors are and After the entire sliding process is completed, the cosine values of the angles corresponding to all candidate areas are counted, and the position corresponding to the maximum cosine value of the angle is selected as the position where the template image and the floating image are optimally matched. This position is the vertical offset relative to the initial search starting point of the template image. This offset is used as the compensation translation amount to perform a longitudinal translation operation on the floating image.

4. The method for automatic whole-eye image stitching based on swept-frequency OCT according to claim 1, characterized in that: The method of using the retinal OCT normal image as a reference image, the retinal OCT left image and the retinal OCT right image as floating images, and using the synchronously acquired laser fundus image to perform horizontal and vertical registration on the retinal image includes: The retinal OCT orthotopic image is used as the reference image, and the retinal OCT left image and the retinal OCT right image are used as floating images. For each of the retinal OCT orthotopic image and the retinal OCT left image or the retinal OCT right image, it corresponds to a laser fundus image. The two laser fundus images are used as the registration input, and the scale-invariant feature transformation algorithm is used to extract the feature points of the laser fundus image. The extracted feature points are matched using a fast nearest neighbor search package to find similar feature point pairs. The matching points are screened using the RANSAC algorithm. Based on the screened valid feature point pairs, the mean of their horizontal distances is calculated to obtain the lateral overlap length of the laser fundus image. Calculate the horizontal overlap length of the retinal image based on the horizontal overlap length of the laser fundus image as follows: ； in, is the image width of the laser fundus image in the registration center, is the image width of the corresponding retinal OCT image, according to the horizontal overlap length of the retinal image , determining the horizontal alignment position of the reference image and the floating image, and adjusting the position of the floating image in the horizontal direction; The column pixel position corresponding to the reference image and the floating image at the splicing point is determined based on the calculated horizontal overlap length. Based on this column of pixels, the vertical displacement difference between the reference image and the floating image is calculated to obtain the vertical translation parameter. The vertical translation transformation operation is performed on the floating image according to the vertical translation parameter to achieve vertical position alignment.

5. The method for automatic whole-eye image stitching based on swept-frequency OCT according to claim 4, characterized in that: The curvature correction of the retinal OCT mosaic image according to the retinal OCT imaging characteristics and the optical structure model to generate a retinal OCT mosaic image close to the actual fundus curvature includes: Curvature correction maps the stitched image into a sector image structure from the posterior end of the vitreous to the fundus, with the center of the sector being the center of the circle. The scanning axis point at the pupil is used to calculate the opening angle corresponding to the complete stitched image based on the stitching width of the image and the optical parameters of OCT imaging. as follows: ； in, is the scanning field angle of the retinal OCT image, is the width of the retinal OCT stitching image, is the image width of the retinal OCT image; For any point in the fan ring structure , from the center of the circle Physical distance The calculation formula is as follows based on the optical path and the refractive index of the medium: ； in, and They are The optical path thickness of the lens and vitreous body at the point position, and are the refractive indices of the lens and vitreous humor, respectively; After the calculation is completed, each pixel in the retinal OCT mosaic image is combined with its relative opening angle The determined scanning angle and the corresponding physical distance , transform from the Cartesian coordinate system to the polar coordinate system with the scanning axis point as the pole to obtain its true spatial position distribution, and then through the inverse transformation from polar coordinates to Cartesian coordinates, map the retinal OCT stitching image back to the plane image to complete the correction of curvature distortion.

6. The method for automatic whole-eye image stitching based on swept-frequency OCT according to claim 1, characterized in that: The step of stitching the anterior segment OCT stitched image and the retinal OCT stitched image to obtain a full-eye image includes: Adjust the size of the retinal OCT mosaic image to keep it consistent with the physical resolution of the anterior segment OCT mosaic image; When stitching images, find the intersection of the visual axis and the line connecting the two anterior chamber angles in the anterior segment OCT stitching image, and determine the relative positions of the anterior segment OCT stitching image and the retinal OCT stitching image based on the coordinates of the intersection; The anterior segment OCT stitching image and the retinal OCT stitching image are moved and stitched according to their relative positions to obtain a whole-eye stitching image.

7. An automatic whole-eye image stitching system based on swept-frequency OCT, characterized by: include: An image acquisition module, configured to acquire anterior segment images and retinal images at different viewing angles in combination with manual guidance; the anterior segment images at different viewing angles include anterior segment OCT anteroposterior image, anterior segment OCT left image, and anterior segment OCT right image; the retinal images at different viewing angles include retinal OCT anteroposterior image and corresponding laser fundus image, retinal OCT left image and corresponding laser fundus image, and retinal OCT right image and corresponding laser fundus image; an anterior segment image stitching module, configured to pre-process the anterior segment images of different viewing angles to obtain anterior chamber angle information, perform rigid registration on the anterior segment images based on the anterior chamber angle information, and fuse the registered anterior segment images to obtain an anterior segment OCT stitching image, including: performing refractive correction on anterior segment OCT anteroposterior images in the anterior segment images; The anterior segment images were preprocessed to obtain a mask image of the effective area, including the cornea and iris. The upper and lower boundary points of the anterior chamber angle were extracted based on the mask image of each anterior segment image, and the registration parameters were obtained through linear fitting. The anterior segment OCT anteroposterior image was used as the reference image, and the left and right anterior segment OCT images were used as floating images. Rigid registration was performed based on the obtained registration parameters. The registered anterior segment images were fused using a grayscale weighted method to generate an anterior segment OCT mosaic image. A retinal image stitching module is used to register retinal images of different viewing angles in the horizontal and vertical directions using synchronously acquired laser fundus images, fuse the registered retinal images, and perform curvature correction on the registered retinal images to obtain a retinal OCT stitched image. The module includes: using the retinal OCT orthotopic image as a reference image, the retinal OCT left image and the retinal OCT right image as floating images, and using the synchronously acquired laser fundus images to perform horizontal and vertical registration on the retinal images; using a grayscale weighted method to fuse the registered retinal images to generate a retinal OCT stitched image; and performing curvature correction on the retinal OCT stitched image based on the retinal OCT imaging characteristics and optical structure model to generate a retinal OCT stitched image that is close to the actual fundus curvature. The whole-eye image stitching module is used to stitch the anterior segment OCT stitching image and the retinal OCT stitching image to obtain a whole-eye image.

8. An electronic device, characterized in that: The device includes a processor and a memory; the memory stores a program, and the program is loaded and executed by the processor to implement the method for automatic whole-eye image stitching based on swept-frequency OCT as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that The storage medium stores a program, which, when executed by the processor, is used to implement the method for automatic whole-eye image stitching based on swept-frequency OCT as described in any one of claims 1 to 6.

Images