ROP characteristic lesion visualization method

Abstract

The invention discloses an ROP characteristic lesion visualization method, which comprises the following steps: acquiring basic parameters of fundus images and basic parameters of ROP characteristic lesions, and splicing the fundus images into an ultra-wide-angle fundus image; respectively obtaining ROP partition information and ROP staging levels through the basic parameters of the eye fundus image and the basic parameters of the ROP characteristic lesions; and performing clock coordinate conversion on the ultra-wide-angle eye fundus image, and outputting a quantized ROP characteristic lesion region. According to the invention, a doctor can know the condition of a lesion area more accurately.

Description

Technical Field

[0001] The present invention relates to the cross - field of clinical ophthalmology and information technology, and particularly to a method for visualizing characteristic lesions of ROP. Background Art

[0002] As a major cause of childhood blindness, ROP is a neurovascular retinal disease characterized by abnormal fibro - vascular proliferation changes at the junction of the vascular and avascular areas of the retina in premature infants.

[0003] In recent years, national policy expenditures have continuously expanded the accessibility of ROP screening. However, the shortage and uneven distribution of relevant screening professionals, equipment, and technologies have led to the fact that ROP screening programs still cannot be fully implemented. In addition, the lack of unified diagnostic and treatment standards during the screening process results in uneven quality of diagnosis and treatment, which has become another major reason hindering the implementation of ROP screening programs. However, by training relevant medical staff through short - term programs, although preliminary screening can be carried out, accurate diagnosis and treatment recommendations cannot be provided, and phenomena such as missed diagnosis and repeated screening often occur, resulting in delayed treatment or over - medical treatment. In the absence of guidance from specialist doctors, grass - roots screening personnel cannot accurately identify characteristic lesions of ROP for ROP diagnosis and treatment, and their willingness to engage in related diagnosis and treatment work is greatly reduced. Against this background, artificial intelligence technology, especially deep - learning technology, has been increasingly widely applied in the medical field. Deep - learning technology can automatically extract image features, perform efficient and accurate image recognition and analysis, providing a new solution for the feature quantification and intelligent diagnosis of ROP. Using deep - learning algorithms to quantitatively analyze characteristic lesions of ROP on fundus color photographs can significantly improve the screening efficiency and diagnostic accuracy, providing strong support for the early intervention and treatment of ROP.

[0004] The prior art mainly focuses on the research and development of ophthalmic intelligent graphic report systems. Eagle Eye Tech (02251.HK), as the first medical AI stock, is a leader in the domestic AI imaging field and also plays a leading and pioneering role in the global AI retinal imaging field. It has built an integrated software - hardware product structure, leading SaMD software, and a portable fundus camera, and has taken the lead in commercializing in relevant in - hospital departments and out - of - hospital large - health scenarios. With the help of retinal image artificial intelligence recognition technology, multiple eye diseases can be detected in one fundus examination. In addition, the full - life - cycle eye health intelligent ophthalmopathy diagnosis and treatment early - warning platform developed by the Shantou International Eye Center realizes the efficient and accurate detection of multiple fundus diseases including ROP by combining a series of imaging examination technologies such as OCT, OCT, fundus images, and B - ultrasounds. However, the graphic reports of existing ROP detection systems lack interpretability and are difficult to be widely promoted and applied on a large scale. Summary of the Invention

[0005] To solve the above technical problems, the present invention proposes a method for visualizing characteristic lesions of ROP, which solves the problems of low efficiency, low accuracy, and poor report interpretability existing in the existing ROP screening process, and provides strong support for the early intervention and treatment of ROP.

[0006] On the one hand, to achieve the above object, the present invention provides a method for visualizing characteristic lesions of ROP, including:

[0007] Obtain the basic parameters of the fundus image and the basic parameters of the characteristic lesions of ROP, and splice the fundus images into an ultra-wide-angle fundus image;

[0008] Obtain the ROP zone information and the ROP staging grade respectively through the basic parameters of the fundus image and the basic parameters of the characteristic lesions of ROP;

[0009] Perform clock coordinate transformation on the ultra-wide-angle fundus image, and output the quantified characteristic lesion area of ROP.

[0010] Preferably, the basic parameters of the fundus image include the segmentation result of the fundus blood vessels and the positions of the optic disc center and the fovea centralis; the basic parameters of the characteristic lesions of ROP include the dividing line between the avascular area and the vascular area, ridge-like lesions, neovascularization, retinal detachment, and the diagnostic information and segmentation results of additional lesions.

[0011] Preferably, splicing the fundus images into an ultra-wide-angle fundus image includes:

[0012] Judge the eye position information of each fundus image according to the relative position of the structural characteristics of the fundus, select a reference image through the eye position information, and register and splice the reference image and the image to be spliced in the order of first the center and then the periphery to obtain the ultra-wide-angle fundus image;

[0013] Among them, the reference image is a fundus map with an optic disc and centered on the macula.

[0014] Preferably, performing the registration and splicing includes:

[0015] Register the reference image and the image to be spliced;

[0016] Find the common feature points of the reference image and the image to be spliced, and calculate the transformation matrix based on the common feature points;

[0017] Calculate the area contributed by each image to the spliced image through the transformation matrix, and perform image splicing and fusion according to the contribution area to generate the spliced image.

[0018] Preferably, the calculation method of the contribution area is:

[0019] Extract the region of interest based on the fundus image;

[0020] Process the region of interest through the transformation matrix to obtain all image quality distribution maps;

[0021] Perform weighted processing according to the image quality distribution and distance information, and assign the pixels in the overlapping region of interest to the corresponding images through the weighted information to obtain the contribution region;

[0022] Among them, the distance information includes the information of the common region from the reference image to the image to be stitched.

[0023] Preferably, obtaining the ROP partition information includes:

[0024] Obtain the ROP partition information through the basic parameters of the fundus image, the basic parameters of the ROP characteristic lesions, and the preset partition rules;

[0025] Among them, the preset partition rules are: Region I is a circular area with the optic disc as the center and a radius twice the distance from the optic disc center to the fovea of the macula; Region II is a circular area with the optic disc as the center and a radius from the optic disc center to the nasal ora serrata; The remaining area outside Region II is Region III.

[0026] Preferably, obtaining the ROP staging grade includes:

[0027] Obtain the ROP staging grade through the basic parameters of the fundus image, the basic parameters of the ROP characteristic lesions, and the preset staging criteria;

[0028] Among them, the preset staging criteria are: Stage 1 is the appearance of a demarcation line between the peripheral avascular area and the terminals of the posterior pole retinal vessels; Stage 2 is the widening and elevation of the demarcation line, with the appearance of a raised ridge-like lesion; Stage 3 is the appearance of new blood vessels on the ridge-like lesion; Stage 4 is the appearance of partial retinal detachment; Stage 5 is the appearance of total retinal detachment; Plus lesion is the tortuous dilation of retinal arteriovenous.

[0029] Preferably, calculating the position of the ROP lesion area in the clock coordinate system includes:

[0030] Taking the optic disc center of the ultra-wide field of view fundus image as the origin, establish a clock coordinate system, divide the periphery of the optic disc into several fan-shaped areas, corresponding to several different moments of the clock;

[0031] According to the boundary position of the ROP lesion area, calculate its position in the clock coordinate system, and quantify the azimuth angle of the ROP lesion area relative to the optic disc center, that is, the corresponding clock point.

[0032] Preferably, quantifying the azimuth angle of the ROP lesion relative to the center of the optic disc includes:

[0033] Determine the center point of the optic disc. Based on the center of the optic disc, determine the target segmentation line associated with the center point of the optic disc. When the target segmentation line intersects the first pixel point of the ROP lesion area, record it as the starting point information of the ROP lesion area. When the last pixel point of the ROP lesion area intersects the target segmentation line, record it as the ending point information of the ROP lesion area. Obtain the azimuth angle of the ROP lesion relative to the center of the optic disc according to the starting point information and the ending point information, and finally convert the azimuth angle into clock positions;

[0034] Among them, the target segmentation line is obtained by rotating along the circumferential direction of the optic disc according to a preset rule with the optic disc as the center.

[0035] On the other hand, to achieve the above object, the present invention also provides a visualization system for ROP characteristic lesions, including:

[0036] A data construction module: used to obtain the basic parameters of the fundus image and the basic parameters of the ROP characteristic lesion, and splice the fundus images into an ultra-wide-angle fundus image;

[0037] A lesion area zoning module: used to zone the ultra-wide-angle fundus image based on the basic parameters of the fundus image and a preset zoning rule to obtain ROP zoning information;

[0038] A lesion area staging module: used to obtain the corresponding ROP staging grade according to the preset staging standard through the basic parameters of the fundus image and the basic parameters of the ROP characteristic lesion;

[0039] A visualization result output module: used to perform clock coordinate conversion on the ultra-wide-angle fundus image and output the visualized result of the quantified ROP characteristic lesion.

[0040] Compared with the prior art, the present invention has the following advantages and technical effects:

[0041] (1) The present invention can very intuitively express the specific position of the ROP lesion area in the fundus, which helps doctors understand the situation of the lesion area more accurately and make a diagnosis and treatment plan accordingly;

[0042] (2) Improve the ROP screening efficiency and diagnostic accuracy: The present invention can automatically extract the characteristic parameters in the fundus image and perform intelligent analysis by using deep learning technology and artificial intelligence algorithms. This automated analysis process is faster and more accurate than traditional manual screening, significantly improving the efficiency and accuracy of ROP screening and reducing the possibility of missed diagnosis and misdiagnosis;

[0043] (3)Enhance the interpretability of the graphic report: By generating a clock coordinate map and a specific location description of the characteristic lesions, the screening results of the present invention are more intuitive and easier to understand. This visualization method provides a clear diagnostic basis for doctors and is also convenient for explaining the condition in doctor-patient communication, solving the problem of poor interpretability of the graphic report in the existing ROP detection system;

[0044] (4)Facilitate ROP screening in primary medical institutions and remote areas: Due to resource and technology limitations, it is often difficult to carry out effective ROP screening work in primary medical institutions and remote areas. The present invention can be deployed in areas with relatively insufficient technology and equipment, reducing the dependence on highly specialized doctors and making ROP screening and diagnosis in remote and resource-poor areas more feasible.

[0045] (5)Facilitate large-scale promotion and application: The automation and visualization characteristics of the present invention endow it with the potential for large-scale application, enabling it to better serve ROP screening projects nationwide and promoting the popularization of ROP prevention and treatment work. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] The drawings constituting a part of this application are used to provide a further understanding of this application. The schematic embodiments of this application and their descriptions are used to explain this application and do not constitute an improper limitation to this application. In the drawings:

[0047] Figure 1 It is a flowchart of a method for visualizing ROP characteristic lesions according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0048] It should be noted that, without conflict, the embodiments in this application and the features in the embodiments can be combined with each other. The following will refer to the drawings and combine the embodiments to detail this application.

[0049] It should be noted that the steps shown in the flowchart of the drawings can be executed in a computer system such as a set of computer executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order than here.

[0050] The present invention proposes a method for visualizing ROP characteristic lesions, such as Figure 1 , including:

[0051] Obtain the basic parameters of the fundus image and the basic parameters of the ROP characteristic lesions, and splice the fundus images into an ultra-wide-angle fundus image;

[0052] Obtain the ROP zone information and the ROP staging grade respectively through the basic parameters of the fundus image and the basic parameters of the ROP characteristic lesions;

[0053] Perform clock coordinate transformation on the ultra-wide-angle fundus image and output the quantized ROP characteristic lesion area.

[0054] It is worth mentioning that by performing clock coordinate transformation on the ultra-wide-angle fundus image based on ROP characteristics and representing the ROP lesion area at the corresponding position in the clock coordinate system, the specific position of the ROP lesion area in the fundus can be expressed very intuitively, which helps doctors understand the situation of the lesion area more accurately and make a diagnosis and treatment plan accordingly. It has the potential for large-scale application, can better serve the ROP screening project nationwide, and promote the popularization of ROP prevention and treatment work.

[0055] Furthermore, in this embodiment, the fundus image is a nine-direction fundus image, including the central direction, temporal direction, superotemporal direction, inferotemporal direction, superior direction, inferior direction, nasal direction, superonasal direction, and inferonasal direction.

[0056] The basic parameters of the fundus image include the segmentation result of the fundus blood vessels and the positions of the optic disc center and the fovea centralis; the basic parameters of the ROP characteristic lesion include the dividing line between the avascular area and the vascular area, ridge-like lesions, neovascularization, retinal detachment, and diagnostic information and segmentation results of additional lesions (Plus lesions). The segmentation results include size, length, area, position, fractal dimension, etc.

[0057] Furthermore, stitching the fundus images into an ultra-wide-angle fundus image includes:

[0058] Judging the eye position information of each image according to the relative position of the structural characteristics of the fundus, selecting a reference image according to the eye position information, and dragging each fundus image into the coordinate position of the reference image in the order of first the center and then the periphery for registration and stitching to obtain the ultra-wide-angle fundus image;

[0059] Among them, the reference image is a fundus image with an optic disc and centered on the macula.

[0060] Specifically, stitch the nine-direction fundus images into an ultra-wide-angle fundus image according to the basic parameters of the fundus image and the ROP characteristic lesion and the preset nine-direction logical relationship.

[0061] The preset nine-direction logical relationship is: the spatial and geometric relationship between the fundus images in the central direction, temporal direction, superotemporal direction, inferotemporal direction, superior direction, inferior direction, nasal direction, superonasal direction, and inferonasal direction, ensuring that the images in each direction can be correctly stitched to form a coherent ultra-wide-angle fundus image.

[0062] The eye position information of each image is determined based on the relative positions of the structural features (optic disc, macula, and vascular arc) of the fundus. A reference image is selected according to the eye position information. The reference image is preferably a fundus image with an optic disc and centered on the macula. After selecting the reference image, each fundus image is dragged into the coordinate position relative to the reference image in the order from the center to the periphery and then registered and stitched.

[0063] Further, performing the registration and stitching includes:

[0064] Performing registration on the reference image and the image to be stitched;

[0065] Finding the common feature points of the reference image and the image to be stitched, and calculating the transformation matrix based on the common feature points;

[0066] Calculating the area contributed by each image to the stitched image through the transformation matrix, and performing image stitching and fusion according to the contribution area to generate the stitched image.

[0067] Specifically, the calculation method of the contribution area is as follows:

[0068] Inputting the fundus image and extracting the ROI region;

[0069] Inputting the transformation matrix of the fundus image, using the transformation matrix to transform the ROI region to the large image coordinates, and calculating all image quality distribution maps;

[0070] Performing weighted processing according to the image quality distribution and distance information, where the distance information includes the information of the common area from the reference image and the image to be stitched, and the closer the distance, the greater the weight value;

[0071] Allocating the pixels of the ROI overlapping region to the corresponding images according to the weighted information to obtain the contribution area.

[0072] Further, obtaining the ROP partition information includes:

[0073] Obtaining the ROP partition information through the basic parameters of the fundus image, the basic parameters of the ROP characteristic lesions, and the preset partition rules;

[0074] Among them, the preset partition rules are: Region I is the area enclosed by a circle with the optic disc as the center and a radius twice the distance from the center of the optic disc to the center of the macula fovea; Region II is the area enclosed by a circle with the optic disc as the center and a radius from the center of the optic disc to the nasal ora serrata; The remaining area outside Region II is Region III.

[0075] Further, obtaining the ROP staging grade includes:

[0076] Obtaining the ROP staging grade through the basic parameters of the fundus image, the basic parameters of the ROP characteristic lesions, and the preset staging criteria;

[0077] Among them, the preset staging criteria are as follows: Stage 1 is the appearance of a demarcation line between the avascular area around and the endings of the retinal blood vessels in the posterior pole; Stage 2 is the widening and heightening of the demarcation line, with the appearance of a raised ridge-like lesion; Stage 3 is the appearance of new blood vessels on the ridge-like lesion; Stage 4 is the appearance of partial retinal detachment; Stage 5 is the appearance of total retinal detachment; Plus lesion is the tortuous dilation of retinal arteriovenous vessels.

[0078] Furthermore, calculate the position of the ROP lesion area in the clock coordinate system, including:

[0079] In this embodiment, taking the optic disc center of the ultra-widefield fundus image as the origin, a clock coordinate system is established, and the periphery of the optic disc is divided into 12 sectors, corresponding to the 12 o'clock positions of the clock;

[0080] According to the boundary position of the ROP lesion area, calculate its position in the clock coordinate system, and quantify the azimuth angle of the ROP lesion relative to the optic disc center, that is, the corresponding clock position.

[0081] Specifically, quantifying the azimuth angle of the ROP lesion relative to the optic disc center includes: determining the optic disc center point, according to the positional relationship between the ROP lesion area and the optic disc center point. That is, taking the optic disc as the center, a coordinate system is established, and the coordinate information of the ROP lesion area is obtained. The coordinate information of the ROP lesion area includes starting point information and ending point information. According to the starting point information and the ending point information, obtain the azimuth angle of the ridge-like lesion relative to the optic disc center, that is, the corresponding clock position.

[0082] Furthermore, quantifying the azimuth angle of the ROP lesion relative to the optic disc center includes:

[0083] Determine the optic disc center point, according to the optic disc center, determine the target segmentation line associated with the optic disc center point. When the target segmentation line intersects with the first pixel point of the ROP lesion area, it is recorded as the starting point information of the ROP lesion area. When the last pixel point of the ROP lesion area intersects with the target segmentation line, it is recorded as the ending point information of the ROP lesion area. According to the starting point information and the ending point information, obtain the azimuth angle of the ROP lesion relative to the optic disc center, and finally convert the azimuth angle into a clock position;

[0084] Among them, the target segmentation line is obtained by rotating circumferentially around the optic disc according to a preset rule.

[0085] Specifically, taking the optic disc center of the ultra-widefield fundus image as the origin, a clock coordinate system is established, and the periphery of the optic disc is divided into 12 sectors, corresponding to the 12 o'clock positions of the clock. That is, perform clock coordinate conversion on the ultra-widefield fundus image, specifically including:

[0086] Convert the fundus image's coordinates. The formula for converting the image coordinates (x, y) to polar coordinates (r, θ) is as follows:

[0087]

[0088] Among them, r is the distance from the origin to the point x, y, and θ is the angle between the coordinate axes.

[0089] Determine adjacent points: Based on the radius and angle of the target point, find adjacent sampling points. These points will be used for subsequent interpolation calculations.

[0090] Calculate the distance and angle differences: For each adjacent point, calculate the distance difference and angle difference between it and the target point. These differences will be used as weight factors in the interpolation operation.

[0091] Bilinear interpolation: Use the distance and angle differences as weights to perform bilinear weighted averaging on the values of adjacent points to obtain the interpolation result of the target point. The specific formula is as follows:

[0092] Distance difference calculation formula: where (x i , y i ) are the coordinates of the adjacent point, and (x 0 , y 0 ) are the coordinates of the target point;

[0093] Angle interpolation calculation formula:

[0094] Final interpolation formula:

[0095] Among them, z 0 is the interpolation result of the target point, z i is the value of the i-th adjacent point, and w i is the weight factor of the i-th adjacent point, calculated based on the distance and angle differences.

[0096] Angle calculation; The angular position and coverage range of the target characteristic lesion area in the polar coordinate image are the clock positions and clock ranges.

[0097] Furthermore, the visualization results of ROP characteristic lesions include the corresponding ROP stage, the partition location where the ROP characteristic lesion is located, the presence or absence of Plus lesions, the clock position where the ROP lesion is located, and the clock coordinate map of the ROP lesion.

[0098] Specifically, an example of the clock position where the characteristic lesion of the output result is located:

[0099] If the output result is "0.5 clock (7 o'clock to 7.5 o'clock)";

[0100] Then 0.5 clock hours: It represents that the lesion area spans a position range of 0.5 clock hours.

[0101] 7 o'clock to 7:30: It represents a clock coordinate system centered on the optic disc, where the lesion area starts at the position of 7 o'clock and ends at the position of 7:30.

[0102] This embodiment also provides a ROP characteristic lesion visualization system, including:

[0103] Data construction module: It is used to obtain the basic parameters of the fundus image and the basic parameters of the ROP characteristic lesion, and splice the fundus images into an ultra-wide-angle fundus image.

[0104] Lesion area partitioning module: It is used to partition the ultra-wide-angle fundus image based on the basic parameters of the fundus image and the preset partitioning rules to obtain ROP partitioning information.

[0105] Lesion area staging module: It is used to obtain the corresponding ROP staging grade according to the preset staging criteria through the basic parameters of the fundus image and the basic parameters of the ROP characteristic lesion.

[0106] Visualization result output module: It is used to perform clock coordinate conversion on the ultra-wide-angle fundus image and output the visualized result of the quantified ROP characteristic lesion.

[0107] The beneficial effects of the present invention are as follows:

[0108] Improve the ROP screening efficiency and diagnostic accuracy: By using deep learning technology and artificial intelligence algorithms, the present invention can automatically extract the characteristic parameters in the fundus image and perform intelligent analysis. This automated analysis process is faster and more accurate than traditional manual screening, significantly improving the efficiency and accuracy of ROP screening and reducing the possibility of missed diagnosis and misdiagnosis.

[0109] Enhance the interpretability of the graphic report: By generating a clock coordinate map and specific position description of the characteristic lesion, the present invention makes the screening result more intuitive and easy to understand. This visualization method provides a clear diagnostic basis for doctors and is also convenient for explaining the condition in doctor-patient communication, solving the problem of poor interpretability of the graphic report in the existing ROP detection system.

[0110] Promote ROP screening in primary medical institutions and remote areas: Due to resource and technical limitations, it is often difficult for primary medical institutions and remote areas to carry out effective ROP screening work. The present invention can be deployed in areas with relatively insufficient technology and equipment, reducing the dependence on highly specialized doctors and making ROP screening and diagnosis in remote and resource-poor areas more feasible.

[0111] Facilitating large-scale popularization and application: The automation and visualization features of the present invention endow it with the potential for large-scale application, enabling it to better serve the ROP screening projects nationwide and promoting the popularization of ROP prevention and treatment work.

[0112] The above is only a preferred specific embodiment of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the technical field of the present application within the technical scope disclosed in the present application should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for visualizing characteristic lesions of ROP, characterized in that: include: Obtain the basic parameters of the fundus image and the basic parameters of the characteristic lesions of ROP, and stitch the fundus images into an ultra-wide-angle fundus image; Respectively obtaining ROP partition information and ROP staging grade through the basic parameters of the fundus image and the basic parameters of the ROP characteristic lesions; The ultra-wide-angle fundus image is converted into clock coordinates, and a quantified ROP characteristic lesion area is output.

2. The method for visualizing characteristic lesions of ROP according to claim 1, characterized in that: The basic parameters of the fundus image include the segmentation results of fundus blood vessels and the positions of the optic disc center and the fovea; the basic parameters of the characteristic lesions of ROP include the boundary between the avascular area and the vascular area, ridge-like lesions, neovascularization, retinal detachment and the diagnostic information and segmentation results of additional lesions.

3. The method for visualizing characteristic lesions of ROP according to claim 1, characterized in that: The fundus images are stitched into an ultra-wide-angle fundus image, comprising: The eye position information of each fundus image is determined according to the relative position of the structural features of the fundus, a reference image is selected according to the eye position information, and the reference image and the image to be spliced ​​are registered and spliced ​​in the order of center first and periphery later to obtain the ultra-wide-angle fundus image; The reference image is a fundus image with an optic disc and centered on the macula.

4. The method for visualizing characteristic lesions of ROP according to claim 3, characterized in that: Performing the registration and stitching includes: Registering the reference image and the image to be stitched; Finding common feature points of the reference image and the image to be stitched, and calculating a transformation matrix based on the common feature points; The area contributed by each image to the stitched image is calculated by the transformation matrix, and the images are stitched and fused according to the contributed areas to generate a stitched image.

5. The method for visualizing characteristic lesions of ROP according to claim 4, characterized in that: The calculation method of the contribution area is: Extracting regions of interest based on fundus images; Processing the region of interest by means of the transformation matrix to obtain all image quality distribution maps; Performing weighted processing according to the image quality distribution and distance information, allocating pixels of the overlapping area of ​​interest to the corresponding image through the weighted information, and obtaining the contribution area; The distance information includes information about the distance between the common area reference image and the image to be stitched.

6. The method for visualizing characteristic lesions of ROP according to claim 1, characterized in that: Obtaining the ROP partition information includes: Acquire the ROP partition information through the basic parameters of the fundus image, the basic parameters of the ROP characteristic lesions and the preset partition rules; Among them, the preset zoning rules are: Zone I is the area with the optic disc as the center and the radius of twice the distance from the center of the optic disc to the fovea; Zone II is the area with the optic disc as the center and the radius of the distance from the center of the optic disc to the nasal serrations; the remaining area outside Zone II is Zone III.

7. The method for visualizing characteristic lesions of ROP according to claim 1, characterized in that: Obtaining the ROP staging level includes: Obtaining the ROP staging grade through the basic parameters of the fundus image, the basic parameters of the ROP characteristic lesions and the preset staging standards; Among them, the preset staging standards are: Stage 1 is the appearance of a dividing line between the peripheral avascular zone and the posterior pole retinal blood vessel endings; Stage 2 is the widening and elevation of the dividing line, and the appearance of raised ridge-like lesions; Stage 3 is the appearance of new blood vessels on the ridge-like lesions; Stage 4 is the occurrence of partial retinal detachment; Stage 5 is the occurrence of complete retinal detachment; Plus lesions are the occurrence of tortuosity and dilation of retinal arteries and veins.

8. The method for visualizing characteristic lesions of ROP according to claim 1, characterized in that: Calculating the position of the ROP lesion area in a clock coordinate system includes: Taking the center of the optic disc of the ultra-wide-angle fundus image as the origin, a clock coordinate system is established, and the periphery of the optic disc is divided into a number of sector-shaped areas corresponding to a number of different times of the clock; According to the boundary position of the ROP lesion area, its position in the clock coordinate system is calculated to quantify the azimuth angle of the ROP lesion area relative to the center of the optic disc, that is, the corresponding clock point position.

9. The method for visualizing characteristic lesions of ROP according to claim 8, characterized in that: Quantify the azimuth of the ROP lesion relative to the center of the optic disc, including: Determine the center point of the optic disc, determine a target segmentation line associated with the center point of the optic disc according to the center of the optic disc, when the target segmentation line intersects with the first pixel point of the ROP lesion area, record it as the starting point information of the ROP lesion area, when the last pixel point of the ROP lesion area intersects with the target segmentation line, record it as the ending point information of the ROP lesion area, obtain the azimuth of the ROP lesion relative to the center of the optic disc according to the starting point information and the ending point information, and finally convert the azimuth into a clock point; The target segmentation line is obtained by rotating along the circumference of the optic disc according to a preset rule with the optic disc as the center.

10. A ROP characteristic lesion visualization system, applied to the ROP characteristic lesion visualization method according to any one of claims 1 to 9, characterized in that: include: Data construction module: used to obtain the basic parameters of fundus images and the basic parameters of ROP characteristic lesions, and stitch the fundus images into ultra-wide-angle fundus images; Lesion area partitioning module: used to partition the ultra-wide-angle fundus image based on the basic parameters of the fundus image and preset partitioning rules to obtain ROP partitioning information; Lesion area staging module: used to obtain the corresponding ROP staging level according to the preset staging standards through the basic parameters of the fundus image and the basic parameters of the characteristic lesions of ROP; Visualization result output module: used for performing clock coordinate conversion on the ultra-wide-angle fundus image and outputting quantized ROP characteristic lesion visualization results.

Images