Film thickness calculation method and device, electronic equipment and storage medium

Abstract

The invention provides a film thickness calculation method and device, electronic equipment and a storage medium, and relates to the technical field of image processing. The film layer thickness calculation method comprises the following steps: determining position coordinates of the macular fovea in a first fundus OCT image, wherein the first fundus OCT image comprises the macular fovea and a to-be-detected tissue layer located on the substrate side of the macular fovea; performing tilt correction on the first fundus OCT image according to the film structure of the to-be-detected tissue layer to obtain a second fundus OCT image; on the basis of the second fundus OCT image, the film thickness of the tissue layer to be measured is calculated according to the position coordinates of the macular central fovea, so that the film thickness calculation error caused by inclination is effectively reduced, and the film thickness calculation accuracy is improved.

Description

Technical Field

[0001] This application relates to the field of image processing technology, specifically to methods, apparatus, electronic devices, and storage media for calculating film thickness. Background Technology

[0002] Optical coherence tomography (OCT) is a technique that uses near-infrared light to perform high-resolution tomographic imaging of tissues, generating cross-sectional images by detecting interference signals of reflected light. This technology has core applications in ophthalmology, with fundus OCT images clearly displaying fine structures such as the retinal nerve fiber layer, macula, and optic disc, making it the "gold standard" for diagnosing macular degeneration, glaucoma, and diabetic retinopathy.

[0003] Due to factors such as inaccurate equipment calibration, patient cooperation, and differences in ocular structure, fundus OCT images may appear tilted. Directly calculating the membrane thickness may introduce errors, thereby increasing the likelihood of clinical misjudgment. Summary of the Invention

[0004] This application provides a method, apparatus, electronic device, and storage medium for calculating film thickness, so as to improve the accuracy of film thickness calculation.

[0005] In a first aspect, embodiments of this application provide a method for calculating membrane thickness, comprising: determining the position coordinates of the fovea centralis in a first fundus OCT image, the first fundus OCT image including the fovea centralis and a tissue layer to be tested located on the basal side of the fovea centralis; performing tilt correction on the first fundus OCT image according to the membrane structure of the tissue layer to be tested to obtain a second fundus OCT image; and calculating the membrane thickness of the tissue layer to be tested based on the second fundus OCT image and the position coordinates of the fovea centralis.

[0006] In conjunction with the first aspect, in one possible implementation, the tilt correction of the first fundus OCT image based on the membrane structure of the tissue layer to be tested to obtain the second fundus OCT image includes: determining a target membrane segment from the tissue layer to be tested whose structural continuity meets the target stability condition; fitting a reference curve of the target membrane segment and calculating the angle between the reference curve and a preset reference line; and tilt correction of the first fundus OCT image based on the angle to obtain the second fundus OCT image.

[0007] In conjunction with the first aspect, in one possible implementation, determining a target membrane segment from the tissue layer to be tested whose structural continuity satisfies the target stability condition includes: determining a local membrane segment located on the basal side of the fovea and whose distance from the fovea is within the target distance threshold from the tissue layer to be tested, and taking the local membrane segment as the target membrane segment.

[0008] In conjunction with the first aspect, in one possible implementation, if the target membrane segment is a local membrane segment in the choroidal membrane, then fitting the reference curve of the target membrane segment includes: fitting the upper boundary of the local membrane segment in the choroidal membrane to obtain the reference curve.

[0009] In conjunction with the first aspect, in one possible implementation, determining the position coordinates of the fovea centralis in the first fundus OCT image includes: performing target detection on the first fundus OCT image to obtain a detection box for the fovea centralis; performing semantic segmentation on the first fundus OCT image to obtain an image region of the retinal layer; and determining the position coordinates of the fovea centralis in the first fundus OCT image based on the detection box and the image region of the retinal layer.

[0010] In conjunction with the first aspect, in one possible implementation, determining the position coordinates of the macular central fovea in the first fundus OCT image based on the detection box of the macular central fovea and the image region of the retinal layer includes: extracting the center line of the detection box of the macular central fovea, the center line being the location of the macular central fovea; determining at least one intersection point between the center line and the image region of the retinal layer, the at least one intersection point having the same abscissa but different ordinate; and determining the position coordinates of the macular central fovea in the first fundus OCT image based on the position coordinates of the at least one intersection point.

[0011] In conjunction with the first aspect, in one possible implementation, the x-coordinate of the fovea is the x-coordinate of at least one intersection point, and the y-coordinate of the fovea is the minimum value of the y-coordinate of at least one intersection point.

[0012] Secondly, embodiments of this application provide a membrane thickness calculation device, comprising: a determining module for determining the position coordinates of the fovea centralis in a first fundus OCT image, the first fundus OCT image including the fovea centralis and a tissue layer to be measured located on the basal side of the fovea centralis; a correcting module for tilting the first fundus OCT image according to the membrane structure of the tissue layer to be measured, thereby obtaining a second fundus OCT image; and a calculation module for calculating the membrane thickness of the tissue layer to be measured based on the second fundus OCT image and the position coordinates of the fovea centralis.

[0013] Thirdly, embodiments of this application provide a computer-readable storage medium storing a computer program for executing the film thickness calculation method mentioned in the first aspect of embodiments of this application.

[0014] Fourthly, embodiments of this application provide an electronic device, which includes: a processor; a memory for storing processor-executable instructions; and a processor for executing the film thickness calculation method mentioned in the first aspect of embodiments of this application.

[0015] The membrane thickness calculation method provided in this application improves the accuracy of membrane thickness calculation by first performing image tilt correction based on the tissue layer to be measured before calculating the membrane thickness. This effectively reduces the error caused by image tilt in membrane thickness calculation. The method also locates the fovea and selects a local membrane segment located on the basal side of the fovea, within a target distance threshold, as a benchmark for membrane correction. Since this local membrane segment typically has fewer undulations and is relatively stable, this method further improves the accuracy of membrane correction. Furthermore, because the membrane thickness at the fovea is crucial for lesion assessment in clinical diagnosis, the membrane thickness calculation method provided in this application, by calculating the membrane thickness of the tissue layer to be measured at the fovea, makes the measurement results more representative and better reflects the lesion condition. This provides a more reliable basis for clinical diagnosis, helps doctors more accurately assess macular degeneration and related diseases, and provides a reliable basis for developing more reasonable treatment plans. Furthermore, images corrected for membrane thickness can improve the accuracy of follow-up examinations, allowing for observation and comparison of changes. Additionally, using the corrected images from this application for model training can enhance the robustness and accuracy of the model. Attached Figure Description

[0016] The above and other objects, features, and advantages of this application will become more apparent from the more detailed description of the embodiments of this application in conjunction with the accompanying drawings. The drawings are provided to further illustrate the embodiments of this application and form part of the specification. They are used together with the embodiments of this application to explain this application and do not constitute a limitation thereof. In the drawings, the same reference numerals generally represent the same components or steps.

[0017] Figure 1 The image shown is a schematic diagram of a fundus OCT image provided in an embodiment of this application.

[0018] Figure 2 This is a flowchart illustrating a film thickness calculation method provided in an embodiment of this application.

[0019] Figure 3 This is a schematic flowchart of a tilt correction method provided in an embodiment of this application.

[0020] Figure 4 This is a schematic flowchart of another tilt correction method provided in the embodiments of this application.

[0021] Figure 5 This is a schematic flowchart of another tilt correction method provided in the embodiments of this application.

[0022] Figure 6This is a flowchart illustrating a method for determining position coordinates provided in an embodiment of this application.

[0023] Figure 7 This is a flowchart illustrating another method for determining position coordinates provided in an embodiment of this application.

[0024] Figure 8 This is a flowchart illustrating another method for calculating film thickness provided in an embodiment of this application.

[0025] Figure 9 This is a schematic diagram of the structure of a film thickness calculation device provided in an embodiment of this application.

[0026] Figure 10 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0027] Hereinafter, exemplary embodiments according to this application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments of this application. It should be understood that this application is not limited to the exemplary embodiments described herein.

[0028] OCT imaging (Optical Coherence Tomography) is a high-resolution medical imaging technique primarily used in ophthalmology, cardiovascular medicine, and dermatology, with ophthalmology being its most widespread application. Its technical principle is similar to ultrasound, but it uses near-infrared light (rather than sound waves) to perform tomographic scanning of tissues. By measuring the interference signal of the reflected light, a cross-sectional image is generated. OCT images can display multilayered membrane structures of different biological tissues or materials. Figure 1 As shown, this embodiment uses fundus OCT images, in which the retinal layer 1 and choroidal layer 2 are clearly displayed from the inside out. The choroidal layer 2 is the vascular layer outside the retina. In OCT retinal imaging, the retinal layer 1 further includes, from the inside out: the internal limiting membrane layer, the retinal nerve fiber layer, the ganglion cell layer, the internal plexiform layer, the nuclear layer, the external plexiform layer, the external nuclear layer, the external membrane layer, the photoreceptor layer, and the retinal pigment epithelium layer. Continuing as... Figure 2 As shown, the fovea 3 is the central depression of the macular region of the retina. It appears as a depression in OCT images and is mainly composed of dense cone cells. It is the core area with the most acute vision.

[0029] Calculating the thickness of the retinal and / or choroidal layers can help doctors diagnose eye diseases more accurately. For example, changes in the thickness of the retinal nerve fiber layer may be associated with diseases such as glaucoma; thinning may indicate glaucoma progression or an increased risk of developing the disease. Changes in choroidal thickness are related to various eye diseases, such as age-related macular degeneration (AMD). Studies have shown that abnormal changes in choroidal thickness may occur in patients with AMD. Accurate measurement of choroidal thickness helps in the early detection of lesions and timely treatment. Furthermore, for patients with diabetic retinopathy, dynamic monitoring of retinal thickness can reflect the stage of disease progression, allowing doctors to adjust treatment plans based on changes in thickness. In clinical practice, accurate calculation of retinal thickness plays an indispensable role in the diagnosis, monitoring, and treatment decisions of ophthalmological diseases. It provides ophthalmologists with quantitative indicators, making diagnosis more precise and treatment more targeted.

[0030] However, due to factors such as inaccurate equipment calibration, patient cooperation, and differences in ocular structure, fundus OCT images may appear tilted. Figure 1 As shown, the tilt mentioned here mainly refers to the angle between the plane containing the membrane and the horizontal direction or the bottom edge of the image. If the membrane thickness is calculated directly based on the thickness of the membrane in the vertical direction in a tilted OCT image, it may cause errors in the membrane thickness, thereby increasing the possibility of clinical misjudgment.

[0031] This embodiment of the application improves the accuracy of membrane thickness calculation by performing image tilt correction based on the tissue layer to be measured before calculating the membrane thickness. Furthermore, this embodiment uses a local membrane segment located on the basal side of the fovea and within a target distance threshold as a reference for membrane correction. Since this local membrane segment typically has fewer undulations and is relatively stable, this embodiment further enhances the accuracy of membrane correction.

[0032] Furthermore, given the crucial role of the membrane thickness at the fovea in clinical diagnosis for assessing lesions, this application's embodiments calculate the membrane thickness of the tissue layer to be measured at the fovea, making the measurement results more representative and accurately reflecting the lesion condition. This method provides a more reliable basis for clinical diagnosis, helping doctors to more accurately assess macular degeneration and related diseases, thereby providing scientific support for developing more rational treatment plans.

[0033] The membrane thickness calculation method provided in this application can be applied to fundus OCT image acquisition devices or analysis terminals connected to such devices. For example, in an ophthalmology hospital examination room, after the fundus OCT device acquires images of the patient, it transmits the image data to the analysis terminal in real time. The analysis terminal uses the membrane thickness calculation method of this application to quickly and accurately calculate the membrane thickness of the retinal layer and choroidal layer, and other tissue layers to be tested. By reviewing the calculation results provided by the analysis terminal and combining them with other clinical symptoms and examination data, doctors can make a more comprehensive and accurate judgment on the patient's eye health.

[0034] In practical applications, the membrane thickness calculation method of this application embodiment can also be integrated into the hospital's electronic medical record system. When a patient's fundus OCT image data is entered into the electronic medical record system, the system automatically calls the calculation method to calculate the membrane thickness and stores the calculation result along with the patient's other examination information. In this way, doctors can easily review the changes in the patient's membrane thickness at different times during subsequent diagnosis and follow-up, observe the development trend of the disease more intuitively, and provide strong data support for long-term disease management.

[0035] The following is an exemplary description of the film thickness calculation method according to embodiments of this application.

[0036] Figure 2 This is a schematic flowchart illustrating a film thickness calculation method provided in an embodiment of this application. See also... Figure 2 The film thickness calculation method provided in this application includes the following steps.

[0037] S10, determine the position coordinates of the fovea centralis in the first fundus OCT image.

[0038] The first fundus OCT image includes the fovea centralis and the tissue layer to be measured located on the basal side of the fovea centralis. The first fundus OCT image is a tilted fundus OCT image. The tissue layer to be measured is the membrane layer whose thickness is to be measured; exemplarily, the tissue layer to be measured includes the retinal layer and / or the choroidal layer, etc.

[0039] Optionally, the position coordinates of the fovea in the first fundus OCT image can be determined by: identifying the fovea from the first fundus OCT image using pre-extracted image features of the fovea; and determining the position coordinates of the fovea in the first fundus OCT image based on the identification result. For example, the position coordinates corresponding to the identification result can be used as the position coordinates of the fovea in the first fundus OCT image.

[0040] S20, the first fundus OCT image is tilted according to the membrane structure of the tissue layer to be tested, and the second fundus OCT image is obtained.

[0041] The second fundus OCT image is the fundus OCT image after tilting has been eliminated.

[0042] Optionally, tilt correction can be performed on the first fundus OCT image based on the entire membrane structure. For example, a curve of the entire membrane structure can be fitted; the tilt angle can be calculated based on the fitted curve; and the tilt angle can be used to correct the tilt of the first fundus OCT image.

[0043] S30, based on the second fundus OCT image, calculates the membrane thickness of the tissue layer to be measured according to the position coordinates of the fovea of ​​the macula.

[0044] The membrane thickness of the tissue layer to be measured is the thickness of a local membrane segment within a set distance threshold from the fovea of ​​the macula, that is, the membrane thickness of the tissue layer to be measured at the fovea of ​​the macula.

[0045] Optionally, the membrane thickness of the tissue layer to be measured is calculated based on the position coordinates of the fovea of ​​the macular region using the second fundus OCT image, including: determining the tilt-corrected position coordinates of the fovea of ​​the macular region based on the position coordinates of the second fundus OCT image; and determining the membrane thickness of the tissue layer to be measured based on the tilt-corrected position coordinates of the fovea of ​​the macular region.

[0046] For example, determining the membrane thickness of the tissue layer to be tested based on the position coordinates of the fovea after tilt correction includes: drawing a perpendicular line along the position coordinates of the fovea after tilt correction, determining the intersection point of the perpendicular line with the upper boundary and the lower boundary of the tissue layer to be tested, and determining the membrane thickness of the tissue layer to be tested based on the distance between the intersection points of the upper and lower boundaries.

[0047] This application does not strictly limit the execution order of S10, S20, and S30. Optionally, S20 can be executed before S10. Preferably, S10 is executed before S20. If the macula is corrected before it is located, its position may be offset due to the correction, resulting in a deviation in the vertical direction. This application preferably locates the fovea centralis first, and then performs coordinate transformation. On the one hand, the fovea centralis is a core landmark of the fundus, and its position determines the selection of the correction reference area. The coordinate transformation must take the initial macula position as the origin to ensure that the thickness calculation position does not shift. On the other hand, it avoids the superposition of errors caused by the deviation in the selection of the reference area or secondary positioning. This application can ensure the measurement accuracy of the fovea centralis, a core diagnostic site.

[0048] The technical solution of this application embodiment, by first performing image tilt correction based on the tissue layer to be measured before calculating the membrane thickness, effectively reduces the error in membrane thickness calculation caused by image tilt, thereby improving the accuracy of membrane thickness calculation. Furthermore, given the important role of membrane thickness at the fovea in clinical diagnosis for lesion assessment, this application embodiment calculates the membrane thickness of the tissue layer to be measured at the fovea, making the measurement results more representative and accurately reflecting the lesion condition. This method provides a more reliable basis for clinical diagnosis, helping doctors to more accurately assess macular degeneration and related diseases, thus providing scientific support for developing more reasonable treatment plans.

[0049] Figure 3 This is a schematic flowchart illustrating a tilt correction method provided in an embodiment of this application. See also... Figure 3 ,exist Figure 2 Based on the disclosed information, S20 may include the following steps.

[0050] S21, determine the target membrane segment whose structural continuity meets the target stability condition from the tissue layer to be tested.

[0051] The target stationarity condition can be: the rate of change of curvature is less than a set threshold for the rate of change in a given region. The target membrane segment is the membrane segment in the tissue layer to be tested that meets the target stationarity condition.

[0052] Optionally, the target membrane segment whose structural continuity satisfies the target stationarity condition is determined from the tissue layer to be tested. This includes: segmenting the tissue layer to be tested to obtain multiple local membrane segments; performing curve fitting on each local membrane segment and determining whether the fitted curve satisfies the target stationarity condition; and selecting the local membrane segment corresponding to the curve that satisfies the target stationarity condition as the target membrane segment. The fitted curve having an angle of no more than 5° with the horizontal line is considered to satisfy the target stationarity condition.

[0053] Optionally, the target membrane segment whose structural continuity meets the target stability condition is determined from the tissue layer to be tested, including: taking the width of the detection frame of the fovea as the length unit, taking the range of N detection frames on the left and right sides of the fovea as the region of the local membrane segment, so as to facilitate curve fitting of the local membrane segment and determine whether the fitted curve meets the target stability condition, where 2≤N.

[0054] In another alternative implementation, the image is divided into M equal parts. Equal portions of the image on either side of the fovea are taken as target membrane segments, for example, the quarter and three-quarters of the image. A curve is fitted to the upper boundary of the choroidal membrane segmented within this region. The angle between the fitted curve and the horizontal line is calculated. Based on this angle, the image is corrected to a horizontal plane using an affine transformation. The thicknesses of the two membrane layers below the fovea are then calculated. This method offers faster detection speeds.

[0055] S22, Fit the reference curve of the target membrane segment and calculate the angle between the reference curve and the preset reference line.

[0056] The angle between the baseline curve and the preset reference line is used to represent the tilt angle of the first fundus OCT image. The preset reference line serves as a reference for calculating the tilt angle of the first fundus OCT image. Optionally, the preset reference line can be a horizontal line, a vertical line, or a reference line with a set angle, etc. Here, the horizontal line is parallel to the bottom edge of the first fundus OCT image.

[0057] S23, the first fundus OCT image is tilted according to the included angle to obtain the second fundus OCT image.

[0058] There are many methods for tilt correction of images based on tilt angle, and the embodiments of this application do not limit this method.

[0059] This application embodiment performs tilt correction by identifying a target membrane segment from the tissue layer to be tested that meets the target stability condition for structural continuity. Because this target membrane segment has few undulations and is relatively stable, the tilt correction is more accurate, thereby effectively improving the accuracy of membrane thickness calculation. Furthermore, this target membrane segment-based correction method has high feasibility and stability in practical applications. In actual clinical settings, different patients have individual differences in their ocular structures, and the method provided in this application embodiment can adapt well to these differences. Even with complex and varied fundus OCT images, tilt correction and membrane thickness calculation can still be accurately performed by identifying a suitable target membrane segment.

[0060] Correction alters the image coordinate system. Subsequent thickness calculations require the use of the "coordinates of the fovea after tilt correction." If correction is performed before locating the fovea, secondary localization is necessary, increasing computational complexity and the risk of error. This application not only eliminates global tilt but also ensures accurate thickness measurement at the fovea.

[0061] Figure 4 This is a schematic flowchart of another tilt correction method provided in an embodiment of this application. See also... Figure 4 ,exist Figure 3 Based on the publicly disclosed information, S21 may include the following steps.

[0062] S211, Identify a local membrane segment located on the basal side of the fovea of ​​the macula from the tissue layer to be tested, and whose distance from the fovea of ​​the macula is within the target distance threshold, and take the local membrane segment as the target membrane segment.

[0063] For example, the target membrane segment can be a local membrane segment closest to the fovea of ​​the macula.

[0064] Since this local membrane segment on the basal side of the fovea of ​​the macula is usually less undulating and more stable, meeting the target stability condition, using it as the target membrane segment for tilt correction can further improve the accuracy of membrane correction.

[0065] Figure 5 This is a schematic flowchart of another tilt correction method provided in an embodiment of this application. See also... Figure 5 ,exist Figure 3 Based on the publicly disclosed information, S22 may include the following steps.

[0066] S221, If ​​the target membrane segment is a local membrane segment in the choroidal membrane, then fit the upper boundary of the local membrane segment in the choroidal membrane to obtain the reference curve, and calculate the angle between the reference curve and the preset reference line.

[0067] The technical solution of this application embodiment obtains a reference curve by fitting the upper boundary of a local membrane segment in the choroidal layer. Because the choroidal layer has unique structural characteristics in fundus OCT images, its upper boundary is relatively clear, and within a local membrane segment that meets the stability condition, it can more stably reflect the image tilt. By accurately fitting this upper boundary, the angle between the obtained reference curve and the preset reference line can more accurately represent the tilt angle of the first fundus OCT image. Prioritizing the localization of the fovea is a prerequisite for tilt correction, and the two have a strict temporal dependency. Reversing the steps will disrupt the local correction logic centered on the fovea, reducing the clinical effectiveness of thickness measurement.

[0068] Figure 6 This is a flowchart illustrating a method for determining position coordinates provided in an embodiment of this application. See also... Figure 6 ,exist Figure 2 Based on the publicly disclosed information, S10 may include the following steps.

[0069] S11, target detection is performed on the first fundus OCT image to obtain the detection box of the fovea of ​​the macula.

[0070] The bounding box for the fovea is used to mark the approximate area of ​​the fovea in an image. Deep learning algorithms, such as convolutional neural networks (CNNs), can be used for object detection. The model is trained on a large number of fundus OCT images with fovea annotations, allowing it to learn the characteristic patterns of the fovea. When processing the first fundus OCT image, the trained model can identify the location of the fovea and generate the corresponding bounding box. The position and size of this bounding box will vary depending on the actual location of the fovea in the image.

[0071] In a preferred embodiment, the upper boundary of the retinal layer is labeled as the internal limiting membrane (IRM), and the lower boundary as the RPE layer. Similarly, the upper boundary of the choroidal layer is labeled as the RPE layer, and the lower boundary as the choroidal great vessel layer. The region between the upper and lower boundaries is used as input to the segmentation model to train the segmentation network, which then segments the retinal and choroidal layers. A macular center detection box is labeled, extending from the retina to the choroid, with the centerline indicating the location of the fovea. This labeled detection box is then used to train the fovea detection network.

[0072] S12, perform semantic segmentation on the first fundus OCT image to obtain the image region of the retinal layer.

[0073] The retinal layer region in the image is separated from the background and other tissues in the first fundus OCT image using semantic segmentation techniques. Semantic segmentation can employ various methods, such as segmentation algorithms based on fully convolutional neural networks (FCNs). This algorithm learns pixel-level feature representations of the retinal layer by training on a large number of fundus OCT images. When processing the first fundus OCT image, the FCN can predict pixel-by-pixel whether a pixel belongs to the retinal layer, thus obtaining a binary mask image of the same size as the original image, where white pixels represent the retinal layer region and black pixels represent other regions.

[0074] S13, Based on the detection frame of the fovea and the image region of the retinal layer, determine the position coordinates of the fovea in the first fundus OCT image.

[0075] For example, determining the position coordinates of the fovea in the first fundus OCT image based on the detection box of the fovea and the image region of the retinal layer includes: finding the overlapping portion between the detection box of the fovea and the image region of the retinal layer, since the fovea is usually located within the retinal layer, this overlapping portion is the precise area where the fovea may exist. Next, within this overlapping region, feature points such as the geometric center or centroid are calculated, and the coordinates of these feature points are used as the position coordinates of the fovea in the first fundus OCT image.

[0076] The technical solution of this application embodiment, by combining the results of object detection and semantic segmentation to determine the location coordinates of the fovea, fully utilizes the advantages of both technologies. Object detection can quickly locate the approximate range of the fovea, while semantic segmentation precisely defines the region of the retinal layer. The combination of the two can more accurately determine the location coordinates of the fovea, laying a solid foundation for subsequent accurate calculation of the membrane thickness. In practical applications, due to significant differences in fundus OCT images from different patients, factors such as image clarity and noise interference can affect the determination of the location coordinates of the fovea. The embodiments of this application, through the complementarity of the two technologies, have a certain degree of robustness to these interfering factors. Even in cases of poor image quality, the location coordinates of the fovea can still be determined as accurately as possible. For example, when the image has a certain degree of blur, the object detection model may not be able to accurately locate the boundary of the fovea, but semantic segmentation, by grasping the overall features of the retinal layer, can, to some extent, compensate for the shortcomings of object detection, helping to more accurately determine the position of the fovea in the retinal layer, and thus determine its location coordinates in the first fundus OCT image.

[0077] Figure 7 This is a flowchart illustrating another method for determining position coordinates provided in an embodiment of this application. See also... Figure 7 ,exist Figure 2 Based on the publicly disclosed information, S13 may include the following steps.

[0078] S131, Extract the center line of the detection frame of the fovea of ​​the macula.

[0079] The center line marks the location of the fovea centralis.

[0080] S132, determine at least one intersection point between the center line and the image region of the retinal layer.

[0081] At least one intersection point has the same x-coordinate but different y-coordinates. The x-coordinate and y-coordinate of the intersection point refer to its image coordinates. The horizontal axis of the image coordinate system is parallel to the horizontal edge of the first fundus OCT image, and the vertical axis is parallel to the vertical edge of the first fundus OCT image. Because the retinal layer is a single image region, the central line intersects with the image region of the retinal layer at multiple points; these multiple intersection points have the same x-coordinate but different y-coordinates.

[0082] S133, determine the position coordinates of the fovea centralis in the first fundus OCT image based on the position coordinates of at least one intersection point.

[0083] In one possible implementation, the x-coordinate of the fovea is the x-coordinate of at least one intersection point, and the y-coordinate of the fovea is the minimum value of the y-coordinates of at least one intersection point. This is because, under normal circumstances, the fovea is located relatively close to the center of the image in the retinal layer, and this method allows for more accurate positioning of the fovea within the retinal layer.

[0084] The technical solution of this application provides a more specific and operable implementation path by extracting the center line of the fovea detection frame and determining its intersection with the retinal layer image region, thereby determining the coordinates of the fovea. This method utilizes the geometric relationship between the detection frame center line and the retinal layer region, enabling a more intuitive determination of the fovea's position. Compared to some other complex calculation methods, this method ensures accuracy while having a relatively small computational load, making it more suitable for rapidly processing large numbers of fundus OCT images in practical applications. For example, in clinical diagnosis in hospitals, doctors need to obtain information related to fundus lesions in a short time; this method can quickly and accurately determine the coordinates of the fovea, saving time for subsequent membrane thickness calculations and disease diagnosis.

[0085] Figure 8 This is a schematic flowchart illustrating another method for calculating film thickness provided in an embodiment of this application. See also... Figure 8 The film thickness calculation method provided in this application includes the following steps.

[0086] S40, based on the trained fovea detection network, extracts the detection box of the fovea of ​​the macular region from the first fundus OCT image.

[0087] The training of the fovea detection network can be described as follows: the sample image is labeled with a fovea detection box, which runs through the retina to the choroid, and the center line of the detection box is the location of the fovea. The labeled sample image is then input into the initial target detection network model for model training, resulting in the trained fovea detection network.

[0088] S50, based on a trained segmentation network, segments the retinal and choroidal layers from the first fundus OCT image.

[0089] The training of the segmentation network can be described as follows: the upper and lower boundaries of the retinal layer are marked on the sample image, with the upper boundary being the internal limiting membrane and the lower boundary being the retinal pigment epithelium; the upper and lower boundaries of the choroidal layer are marked on the sample image, with the upper boundary being the retinal pigment epithelium and the lower boundary being the choroidal great vessels; the middle region of the upper and lower boundaries is taken and input into the initial segmentation network model for model training, resulting in the trained segmentation network.

[0090] S60, determine the position coordinates of the fovea based on the coordinates of the intersection point between the center line of the detection frame of the fovea and the retinal layer.

[0091] The x-coordinate of the intersection point is the x-coordinate of the fovea, and the minimum y-coordinate of the intersection point is the y-coordinate of the fovea.

[0092] S70, Take the image region including the fovea of ​​the macula from the first fundus OCT image, and fit the upper boundary of the choroid layer obtained by segmenting the image region into a curve.

[0093] For example, the region between a first proportional position and a second proportional position in the first fundus OCT image is selected in the horizontal direction as the image region including the fovea of ​​the macula. The first proportional position is smaller than the second proportional position, for example, the first proportional position is one-third and the second proportional position is two-thirds, etc.

[0094] S80, calculate the angle between the curve and the horizontal line, and perform tilt correction on the first fundus OCT image based on the angle to obtain the second fundus OCT image.

[0095] For example, an affine transformation is performed on the first fundus OCT image based on the included angle to correct the image to a horizontal position.

[0096] S90, based on the second fundus OCT image, calculate the thickness of the local membrane segment on the basal side of the macula, where the distance from the macula is within the target distance threshold, according to the position coordinates of the macula.

[0097] The execution order of S40, S50, S60, S70, and S80 is not limited in this embodiment. Optionally, S50 may be executed before S40, and S70 and S80 may be executed before S60.

[0098] The technical solution of this application achieves significant technical effects through the following innovations: First, it employs multi-layer annotation and segmentation network technology to accurately segment and locate the retinal and choroidal layers, greatly improving the accuracy of macular center localization and model segmentation. Second, it enhances the robustness of the model through multi-feature fusion and foveal detection box design, enabling it to adapt to the detection needs of different types of OCT images. Simultaneously, the fully automated segmentation and detection process significantly reduces the need for manual intervention. Furthermore, this solution has broad applicability, processing various normal and pathological OCT images and supporting the analysis of multiple retinal structures. Most importantly, its high-precision localization capability provides reliable technical support for the early diagnosis and precise treatment of retinal diseases, effectively improving clinical treatment outcomes.

[0099] Figure 9 This is a schematic diagram of a film thickness calculation device provided in an embodiment of this application. See also... Figure 9 The film thickness calculation device provided in this application embodiment includes a determination module 100, a correction module 200, and a calculation module 300.

[0100] The determination module 100 is used to determine the position coordinates of the fovea centralis in the first fundus OCT image, which includes the fovea centralis and the tissue layer to be tested located on the basal side of the fovea centralis; the correction module 200 is used to perform tilt correction on the first fundus OCT image according to the membrane structure of the tissue layer to be tested, to obtain a second fundus OCT image; the calculation module 300 is used to calculate the membrane thickness of the tissue layer to be tested based on the second fundus OCT image and the position coordinates of the fovea centralis.

[0101] In one possible implementation, the correction module 200 is further configured to: determine a target membrane segment whose structural continuity meets the target stability condition from the tissue layer to be tested; fit a reference curve of the target membrane segment and calculate the angle between the reference curve and a preset reference line; and perform tilt correction on the first fundus OCT image based on the angle to obtain a second fundus OCT image.

[0102] In one possible implementation, the correction module 200 is further configured to: determine a local membrane segment located on the basal side of the fovea and within a target distance threshold from the tissue layer to be tested, and use the local membrane segment as the target membrane segment.

[0103] In one possible implementation, if the target membrane segment is a local membrane segment in the choroidal membrane, the correction module 200 is further used to: fit the upper boundary of the local membrane segment in the choroidal membrane to obtain a reference curve.

[0104] In one possible implementation, the determining module 100 is further configured to: perform target detection on the first fundus OCT image to obtain a detection box of the fovea; perform semantic segmentation on the first fundus OCT image to obtain an image region of the retinal layer; and determine the position coordinates of the fovea in the first fundus OCT image based on the detection box of the fovea and the image region of the retinal layer.

[0105] In one possible implementation, the determining module 100 is further configured to: extract the center line of the detection box of the fovea, the center line being the location of the fovea; determine at least one intersection point between the center line and the image region of the retinal layer, the at least one intersection point having the same abscissa but different ordinate; and determine the position coordinates of the fovea in the first fundus OCT image based on the position coordinates of the at least one intersection point.

[0106] In one possible implementation, the x-coordinate of the fovea is the x-coordinate of at least one intersection point, and the y-coordinate of the fovea is the minimum value of the y-coordinate of at least one intersection point.

[0107] This application's embodiments achieve automated and high-precision film thickness calculation through the collaborative work of a determination module, a correction module, and a calculation module. The determination module accurately determines the coordinates of the fovea centralis using a combination of target detection and semantic segmentation, providing a crucial foundation for subsequent calculations. The correction module performs tilt correction on the image based on the film structure of the tissue layer under test, effectively improving the image's accuracy and usability. The calculation module then uses the corrected image and the determined fovea centralis coordinates to accurately calculate the film thickness of the tissue layer under test. This modular design not only improves computational efficiency but also enhances the scalability and adaptability of the entire system.

[0108] Below, for reference Figure 10 This describes an electronic device according to embodiments of the present application. Figure 10 The diagram shown is a structural schematic of an electronic device provided in an exemplary embodiment of this application.

[0109] like Figure 10 As shown, the electronic device 1000 includes one or more processors 1001 and memory 1002.

[0110] The processor 1001 may be a central processing unit (CPU) or other form of processing unit with data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device 1000 to perform desired functions.

[0111] The memory 1002 may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, random access memory (RAM) and / or cache memory. The non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 1001 may execute the program instructions to implement the film thickness calculation methods of the various embodiments of this application described above and / or other desired functions. Various contents, such as fundus OCT images, may also be stored in the computer-readable storage medium.

[0112] In one example, the electronic device 1000 may also include an input device 1003 and an output device 1004, which are interconnected via a bus system and / or other forms of connection mechanism (not shown).

[0113] The input device 1003 may include, for example, a keyboard, a mouse, etc.

[0114] The output device 1004 can output various information to the outside, including fundus OCT images before and after correction, the position coordinates of the fovea in the first fundus OCT image, and the membrane thickness of the tissue layer to be measured. The output device 1004 may include, for example, a display, a speaker, a printer, and a communication network and its connected remote output devices, etc.

[0115] Of course, for the sake of simplicity, Figure 10 Only some of the components of the electronic device 1000 relevant to this application are shown in this illustration; components such as buses, input / output interfaces, etc., are omitted. In addition, the electronic device 1000 may include any other suitable components depending on the specific application.

[0116] In addition to the methods and apparatus described above, embodiments of this application may also be computer program products, which include computer program instructions that, when executed by a processor, cause the processor to perform the steps in the film thickness calculation methods according to various embodiments of this application described above.

[0117] The computer program product can be written in any combination of one or more programming languages ​​to perform the operations of the embodiments of this application. The programming languages ​​include object-oriented programming languages ​​such as Java and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0118] Furthermore, embodiments of this application may also be computer-readable storage media storing computer program instructions thereon, which, when executed by a processor, cause the processor to perform the steps in the film thickness calculation methods according to various embodiments of this application described above.

[0119] The computer-readable storage medium may be any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may, for example, include, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses, or devices, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: electrical connections having one or more wires, portable disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0120] The basic principles of this application have been described above with reference to specific embodiments. However, it should be noted that the advantages, benefits, and effects mentioned in this application are merely examples and not limitations, and should not be considered as essential features of each embodiment of this application. Furthermore, the specific details disclosed above are for illustrative and facilitative purposes only, and are not limitations. These details do not limit the application to the necessity of employing the aforementioned specific details for implementation.

[0121] The block diagrams of devices, apparatuses, devices, and systems involved in this application are merely illustrative examples and are not intended to require or imply that they must be connected, arranged, or configured in the manner shown in the block diagrams. As those skilled in the art will recognize, these devices, apparatuses, devices, and systems can be connected, arranged, and configured in any manner. Words such as “comprising,” “including,” “having,” etc., are open-ended terms meaning “including but not limited to,” and are used interchangeably with them. The terms “or” and “and” as used herein refer to the terms “and / or,” and are used interchangeably with them unless the context clearly indicates otherwise. The term “such as” as used herein refers to the phrase “such as but not limited to,” and is used interchangeably with it.

[0122] It should also be noted that in the apparatus, equipment, and methods of this application, the components or steps can be disassembled and / or recombined. These disassemblies and / or recombinations should be considered as equivalent solutions of this application.

[0123] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use this application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein can be applied to other aspects without departing from the scope of this application. Therefore, this application is not intended to be limited to the aspects shown herein, but rather to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0124] The above description has been given for purposes of illustration and description. Furthermore, this description is not intended to limit the embodiments of this application to the forms disclosed herein. Although numerous exemplary aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations thereof.

Claims

1. A method for calculating film thickness, characterized in that, include: Determine the position coordinates of the fovea centralis in a first fundus OCT image, the first fundus OCT image including the fovea centralis and the tissue layer to be tested located on the basal side of the fovea centralis; The first fundus OCT image is tilted and corrected according to the membrane structure of the tissue layer to be tested to obtain the second fundus OCT image. Based on the second fundus OCT image, the thickness of the membrane layer at a specific location of the tissue layer to be tested is calculated according to the position coordinates of the fovea of ​​the macula.

2. The method according to claim 1, characterized in that, The step of tilt correction of the first fundus OCT image based on the membrane structure of the tissue layer to be tested to obtain a second fundus OCT image includes: Determine the target membrane segment whose structural continuity satisfies the target stability condition from the tissue layer to be tested; Fit the reference curve of the target membrane segment and calculate the angle between the reference curve and the preset reference line; The first fundus OCT image is tilted and corrected according to the included angle to obtain the second fundus OCT image.

3. The method according to claim 2, characterized in that, The step of determining the target membrane segment whose structural continuity satisfies the target stability condition from the tissue layer to be tested includes: A local membrane segment located on the basal side of the fovea and within a target distance threshold from the fovea is identified from the tissue layer to be tested, and this local membrane segment is taken as the target membrane segment.

4. The method according to claim 2, characterized in that, If the target membrane segment is a local segment within the choroidal membrane, then the fitting of the reference curve for the target membrane segment includes: The upper boundary of a local membrane segment in the choroidal membrane layer is fitted to obtain a baseline curve.

5. The method according to claim 1, characterized in that, Determining the position coordinates of the fovea centralis in the first fundus OCT image includes: Target detection is performed on the first fundus OCT image to obtain the detection box of the fovea of ​​the macula; Semantic segmentation is performed on the first fundus OCT image to obtain the image region of the retinal layer; Based on the detection frame of the fovea and the image region of the retinal layer, the position coordinates of the fovea in the first fundus OCT image are determined.

6. The method according to claim 5, characterized in that, The determination of the position coordinates of the fovea in the first fundus OCT image based on the detection frame of the fovea and the image region of the retinal layer includes: Extract the center line of the detection frame of the fovea of ​​the macula, where the center line is the location of the fovea of ​​the macula; Determine at least one intersection point between the center line and the image region of the retinal layer, wherein the at least one intersection point has the same x-coordinate but different y-coordinates; The position coordinates of the fovea of ​​the macula in the first fundus OCT image are determined based on the position coordinates of the at least one intersection point.

7. The method according to claim 6, characterized in that, The x-coordinate of the fovea is the x-coordinate of the at least one intersection point, and the y-coordinate of the fovea is the minimum value of the y-coordinate of the at least one intersection point.

8. A film thickness calculation device, characterized in that, include: The determination module is used to determine the position coordinates of the fovea centralis in a first fundus OCT image, the first fundus OCT image including the fovea centralis and the tissue layer to be tested located on the basal side of the fovea centralis; The correction module is used to perform tilt correction on the first fundus OCT image according to the membrane structure of the tissue layer to be tested, so as to obtain a second fundus OCT image. The calculation module is used to calculate the membrane thickness of the tissue layer to be tested based on the second fundus OCT image and the position coordinates of the fovea of ​​the macula.

9. A computer-readable storage medium storing a computer program for performing the film thickness calculation method according to any one of claims 1 to 7.

10. An electronic device, the electronic device comprising: processor; Memory used to store the processor's executable instructions; The processor is used to execute the film thickness calculation method according to any one of claims 1 to 7.

Images