Ophthalmologic operation navigation method

Abstract

The invention provides an ophthalmologic operation navigation method which comprises the following steps: acquiring operation images and biological parameters of two eyes of a patient, determining a reference image, configuring reference navigation parameters corresponding to the reference image, and constructing an eyeball iris segmentation model; inputting the current frame of operation image and the reference image into an eyeball iris segmentation model, and performing iris fitting on an iris segmentation result; if the current frame of operation image is in an incision stage or an astigmatism axis implantation stage, performing image registration calculation on the current frame of operation image and a reference image according to an iris fitting result to obtain a rotation angle of the current frame of operation image relative to the reference image; calculating a target incision angle or a target astigmatism axis implantation angle based on the rotation angle and the reference navigation parameters; and if the current frame of operation image is in the capsulorhexis stage, calculating the pixel radius of the target capsulorhexis circle based on the iris fitting result and the reference navigation parameters. According to the invention, the accuracy and safety in the operation process are ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of medical image processing, and in particular to an ophthalmic surgery navigation method. Background Art

[0002] The principle of cataract surgery is to remove the cloudy lens in the eye and implant an artificial transparent lens so that light can pass through the lens without obstruction, thus restoring the patient's vision. Incision, capsulotomy and artificial lens are the key steps in cataract surgery, and the accuracy of the incision position, capsulotomy range and artificial lens implantation angle directly affects the degree of postoperative vision recovery of the patient. Currently, the commonly used method is to manually perform interventional coloring on the eyeball under a slit lamp microscope before surgery to assist the doctor in locating key information such as position and angle during surgery. However, the accuracy of this method is greatly affected by the physiological tremor of the start and the size of the coloring point, and it is easy to cause physiological discomfort to the patient.

[0003] To solve this problem, the existing technology uses computers to assist clinicians in obtaining accurate navigation information such as the incision position, the range of the capsulorhexis, and the angle of the artificial lens implantation, so as to improve the accuracy and safety of the operation. Specifically, an image registration method based on iris blood vessels is used to obtain the rotation information of the eyeball, and the navigation information of the binocular image is processed by point cloud according to the calibrated binocular camera system, and then displayed by projecting it to an external display or a spectroscope projection eyepiece. However, for the application scenario of cataract surgery, the existing navigation technology is difficult to effectively solve complex interference situations such as instrument occlusion during surgery and deformation of iris blood vessels caused by surgical operations, and the stability and accuracy of navigation cannot be guaranteed. The 2D observation in the existing navigation observation method cannot meet the actual use needs of clinicians, and the 3D navigation observation based on point cloud information processing has the problem that real-time and high quality cannot be taken into account at the same time.

[0004] In view of this, it is necessary to improve the ophthalmic surgery navigation method in the prior art to solve the above problems. Summary of the invention

[0005] The purpose of the present invention is to solve the problem that the existing ophthalmic surgical navigation methods are difficult to effectively solve complex interference situations such as instrument occlusion during surgery and iris vascular deformation caused by surgical operations, and cannot guarantee the stability and accuracy of navigation. The existing 2D observation cannot meet the actual usage needs of clinicians, and the 3D navigation observation based on point cloud information processing has the problem that real-time and high quality cannot be taken into account at the same time.

[0006] To achieve the above object, the present invention provides an ophthalmic surgery navigation method, comprising:

[0007] Obtain surgical images and biological parameters of both eyes of the patient, determine a reference image, configure reference navigation parameters corresponding to the reference image according to the biological parameters, and construct an eyeball iris segmentation model based on the surgical images;

[0008] Inputting the current frame surgical image and the reference image in the surgical image into the eyeball iris segmentation model, outputting iris segmentation results corresponding to the current frame surgical image and the reference image respectively, and performing iris fitting on the iris segmentation results to obtain iris fitting results;

[0009] If the current frame of the surgical image is in the incision stage or the astigmatism axis implantation stage, perform image registration calculation on the current frame of the surgical image and the reference image according to the iris fitting result to obtain a rotation angle of the current frame of the surgical image relative to the reference image, and calculate a target incision angle or a target astigmatism axis implantation angle based on the rotation angle and the reference navigation parameters;

[0010] If the current frame of the surgical image is in the capsulorhexis stage, the pixel radius of the target capsulorhexis circle is calculated based on the iris fitting result and the reference navigation parameters.

[0011] As a further improvement of the present invention, the reference image is defined by a surgical image or a preoperative examination image captured from the surgical image.

[0012] As a further improvement of the present invention, performing iris fitting on the iris segmentation result to obtain an iris fitting result includes:

[0013] If the current frame surgical image is in the incision stage or the astigmatism axis implantation stage, an iris ellipse fitting is performed on the iris segmentation results corresponding to the current frame surgical image and the reference image respectively, to obtain an iris fitting result consisting of the center coordinates, the major axis pixel radius and the minor axis pixel radius;

[0014] If the current frame surgical image is in the capsulorhexis stage, iris circle fitting is performed on the iris segmentation results corresponding to the current frame surgical image and the reference image respectively to obtain an iris fitting result consisting of circle center coordinates and pixel radius.

[0015] As a further improvement of the present invention, performing image registration calculation on the current frame surgical image and the reference image according to the iris fitting result to obtain a rotation angle of the current frame surgical image relative to the reference image includes:

[0016] Based on the iris fitting result, the current frame surgical image and the reference image are respectively cropped to obtain an image to be registered and a registration reference image;

[0017] Performing a fusion calculation on the image to be registered and the registration reference image to obtain a matching point set consisting of multiple groups of matching points, and screening the matching points in the matching point set according to the iris fitting result to obtain a target matching point set;

[0018] The rotation angle sequence corresponding to each group of matching points in the target matching point set is calculated, and the rotation angle of the current frame surgical image relative to the reference image is calculated based on the rotation angle sequence.

[0019] As a further improvement of the present invention, the method of respectively cropping the current frame surgical image and the reference image based on the iris fitting result to obtain the image to be registered and the registered reference image includes:

[0020] With the center coordinates of the circle as the center point, crop the current frame surgical image and the reference image with a width and height of 2r * ×k(r * ∈{r l ,r s}) as the image to be registered and the registration reference image respectively;

[0021] Where k is the cropping ratio, and k>1, r l refers to the major axis pixel radius, r s refers to the minor axis pixel radius.

[0022] As a further improvement of the present invention, the step of screening the matching points in the matching point set according to the iris fitting result to obtain a target matching point set includes:

[0023] The matching points in the circular area formed by the coordinates of the center of the circle and the radius of the short axis pixel are retained, and each group of matching points in the circular area constitutes the target matching point set.

[0024] As a further improvement of the present invention, the step of calculating the rotation angle of the current frame surgical image relative to the reference image based on the rotation angle sequence includes:

[0025] Calculating the median of the rotation angle sequence, setting a threshold, and eliminating the rotation angle sequence whose difference with the median is greater than the threshold, to obtain a corrected rotation angle sequence;

[0026] The median of the corrected rotation angle sequence or the mean of the corrected rotation angle sequence is used as the rotation angle.

[0027] As a further improvement of the present invention, the reference navigation parameters include: incision angle Physical diameter of capsulorhexis circle R gt and implantation angle

[0028] The target incision angle Target astigmatism axis implantation angle The target capsulorhexis circle pixel radius refers to the rotation angle, and WTW refers to the physical diameter of the irises of both eyes of the patient.

[0029] As a further improvement of the present invention, the ophthalmic surgery navigation method also includes: drawing the current frame surgical image based on the target incision angle or the target astigmatism axis implantation angle or the target capsulorhexis circle pixel radius, and performing 3D interlaced rendering on the drawn current frame surgical image, and inputting the 3D interlaced rendered current frame surgical image into a display device.

[0030] As a further improvement of the present invention, the surgical image is in a 3D left-right format or a 3D top-bottom format, and the current frame surgical image includes: a first current frame surgical image and a second current frame surgical image;

[0031] The drawing of the current frame surgical image based on the target incision angle or the target astigmatism axis implantation angle or the target capsulorhexis circle pixel radius includes:

[0032] Calculating target navigation parameters corresponding to the first current frame surgical image, and drawing the first current frame surgical image and the second current frame surgical image respectively based on the target navigation parameters; or calculating target navigation parameters corresponding to the first current frame surgical image and the second current frame surgical image respectively, and drawing the first current frame surgical image and the second current frame surgical image respectively based on the target navigation parameters corresponding to the first current frame surgical image and the second current frame surgical image respectively;

[0033] Wherein, the target navigation parameters include: the target incision angle or the target astigmatism axis implantation angle or the target capsulorhexis circle pixel radius.

[0034] Compared with the prior art, the present invention has the following beneficial effects:

[0035] Obtain surgical images and biological parameters of both eyes of the patient, determine the reference image, configure the reference navigation parameters corresponding to the reference image according to the biological parameters, and build an eye iris segmentation model based on the surgical image; input the current frame surgical image and the reference image in the surgical image into the eye iris segmentation model, output the iris segmentation results corresponding to the current frame surgical image and the reference image respectively, perform iris fitting on the iris segmentation result, and obtain the iris fitting result; if the current frame surgical image is in the incision stage or the astigmatism axis implantation stage, perform image registration calculation on the current frame surgical image and the reference image according to the iris fitting result, obtain the rotation angle of the current frame surgical image relative to the reference image, and calculate the target incision angle or the target astigmatism axis implantation angle based on the rotation angle and the reference navigation parameters; if the current frame surgical image is in the capsulorhexis stage, calculate the pixel radius of the target capsulorhexis circle based on the iris fitting result and the reference navigation parameters. Therefore, during the incision stage or the astigmatism axis implantation stage, image registration calculation is performed on the current frame surgical image and the reference image based on the iris fitting result. Compared with the image registration method based on iris blood vessels in the prior art, it is not limited by the local area information of iris blood vessels, while ensuring the accuracy of the final target navigation parameters, thereby solving the problem that the existing ophthalmic surgical navigation methods are difficult to effectively solve complex interference situations such as intraoperative instrument occlusion and iris blood vessel deformation caused by surgical operations, and cannot guarantee the stability and accuracy of navigation. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 This is a schematic diagram of the steps of the ophthalmic surgery navigation method shown in the present invention;

[0037] Figure 2 A schematic diagram of specific steps for performing image registration calculation on a current frame surgical image and a reference image according to an iris fitting result to obtain a rotation angle of the current frame surgical image relative to the reference image;

[0038] Figure 3 Schematic diagram for determining the target matching point set;

[0039] Figure 4 The figure is a schematic diagram of specific steps for calculating the rotation angle of the current frame surgical image relative to the reference image based on the rotation angle sequence. DETAILED DESCRIPTION

[0040] The present invention is described in detail below in conjunction with the various embodiments shown in the accompanying drawings, but it should be noted that these embodiments are not limitations of the present invention, and any equivalent transformations or substitutions in functions, methods, or structures made by ordinary technicians in the field based on these embodiments are all within the scope of protection of the present invention.

[0041] Please refer to Figures 1 to 4As shown, the present invention shows a specific implementation of an ophthalmic surgical navigation method, which is used in cataract surgery to assist clinicians in obtaining accurate target navigation parameters such as target incision angle, target astigmatism axis implantation angle, and target capsulorhexis circle pixel radius, so as to improve the accuracy and safety during the operation.

[0042] Ginseng Figure 1 As shown, the ophthalmic surgery navigation method includes the following steps S1 to S4.

[0043] Step S1, obtaining surgical images and biological parameters of both eyes of the patient, and determining a reference image, configuring reference navigation parameters corresponding to the reference image according to the biological parameters, and constructing an eyeball iris segmentation model based on the surgical images.

[0044] Specifically, before the operation, the measuring instrument collects the preoperative examination images and biological parameters of the patient's eyes; during the operation, the imaging system collects the surgical images of the patient's eyes in real time. Among them, the measuring instrument can be a device for measuring the biological characteristics of the eyeball, such as a biometer, and the imaging system can be a whole composed of an ophthalmic surgical microscope and a 3D microsurgery camera. As long as the collection of preoperative examination images, biological parameters and surgical images can be achieved, the specific equipment in this embodiment is not limited. The surgical images collected by the imaging system are in 3D left-right format or 3D top-down format, and accordingly, the preoperative examination images are also in 3D left-right format or 3D top-down format, so as to facilitate the subsequent processing of the images accordingly. Based on this, the reference image is defined by the surgical image or the preoperative examination image captured in the surgical image, that is, the surgical image captured in the surgical image can be used as the reference image, or the preoperative examination image can be used as the reference image. The biological parameters include: one or any combination of the patient's corresponding axial length, angle curvature, corneal thickness, anterior chamber depth, lens thickness, vitreous thickness, pupil diameter, and white-to-white distance. After collecting the biological parameters, the reference navigation parameters corresponding to the reference image are configured according to the biological parameters. The reference navigation parameters include: incision angle Physical diameter of capsulorhexis circle R gt and implantation angle The specific configuration method can be any method in the prior art, and this embodiment does not specifically limit this. After the surgical images are collected, an eye iris segmentation model is constructed based on the surgical images, including: constructing an iris segmentation data set based on the collected surgical images, designing an iris segmentation network, the network adopts a resnet18+fcn architecture, and training the model to obtain an eye iris segmentation model. The reasoning prediction results of the trained eye iris segmentation model on the test set are compared with the manually annotated gold standard, and the accuracy is 97.04% evaluated by miou, where miou represents the degree of overlap between the model reasoning results and the gold standard results. The higher the value, the more accurate the model.

[0045] Step S2: input the current frame surgical image and the reference image in the surgical image into the eye iris segmentation model, output the iris segmentation results corresponding to the current frame surgical image and the reference image respectively, perform iris fitting on the iris segmentation results, and obtain iris fitting results.

[0046] Specifically, during the operation, the current frame surgical image in the surgical image is acquired in real time, and the current frame surgical image and the reference image are input into the eye iris segmentation model, and the iris segmentation results corresponding to the current frame surgical image and the reference image are output respectively. Then, the iris segmentation results are iris fitted according to the stage of the current frame surgical image to obtain the iris fitting result.

[0047] It should be noted that cataract surgery specifically includes three stages: incision, capsulotomy and astigmatism axis implantation. Different target navigation parameters are required for each stage. For example, the target incision angle needs to be calculated in the incision stage, the target capsulotomy circle pixel radius needs to be calculated in the capsulotomy stage, and the target astigmatism axis implantation angle needs to be calculated in the astigmatism axis stage. In addition, each stage requires different processing of the current frame surgical image and the reference image. Therefore, the iris segmentation result needs to be processed accordingly according to the stage of the current frame surgical image.

[0048] In one embodiment, iris fitting is performed on the iris segmentation result to obtain the iris fitting result, including: if the current frame of the surgical image is in the incision stage or the astigmatism axis implantation stage, iris ellipse fitting is performed on the iris segmentation results corresponding to the current frame of the surgical image and the reference image respectively, to obtain the iris fitting result consisting of the center coordinates, the major axis pixel radius and the minor axis pixel radius; if the current frame of the surgical image is in the capsulorhexis stage, iris circle fitting is performed on the iris segmentation results corresponding to the current frame of the surgical image and the reference image respectively, to obtain the iris fitting result consisting of the center coordinates and the pixel radius. Due to the different characteristics of different stages, the eyeball is less squeezed in the capsulorhexis stage, so iris circle fitting is used to perform iris fitting in the capsulorhexis stage.

[0049] Step S3: If the current frame surgical image is in the incision stage or the astigmatism axis implantation stage, image registration calculation is performed on the current frame surgical image and the reference image according to the iris fitting result to obtain the rotation angle of the current frame surgical image relative to the reference image, and the target incision angle or the target astigmatism axis implantation angle is calculated based on the rotation angle and the reference navigation parameters.

[0050] In one embodiment, Figure 2 As shown, if the current frame surgical image is in the incision stage or the astigmatism axis implantation stage, the current frame surgical image and the reference image are image registered according to the iris fitting result to obtain the rotation angle of the current frame surgical image relative to the reference image, which specifically includes the following steps S31 to S33.

[0051] Step S31 : cropping the current frame surgical image and the reference image respectively based on the iris fitting result to obtain the image to be registered and the registration reference image.

[0052] Specifically, the center coordinates of the circle in the iris fitting result are used as the center point, and the width and height of the current frame surgical image and the reference image are cropped. * ×k(r * ∈{r l ,r s}) as the image to be registered and the registration reference image respectively; where k is the cropping ratio, and k>1, r l is the major axis pixel radius, r s is the minor axis pixel radius.

[0053] Step S32: performing fusion calculation on the image to be registered and the registration reference image to obtain a matching point set consisting of multiple groups of matching points, and screening the matching points in the matching point set according to the iris fitting result to obtain a target matching point set.

[0054] Specifically, a global perception multi-scale registration fusion algorithm is constructed, and based on the algorithm, a fusion calculation is performed on the image to be registered and the registration reference image to obtain a matching point set composed of multiple groups of matching points. Since the global perception multi-scale registration fusion algorithm belongs to the prior art, its specific implementation is not described in detail in this embodiment. The matching points in the matching point set are screened according to the iris fitting result to obtain a target matching point set, including: retaining the matching points in the center area of ​​the circle composed of the center coordinates and the short axis pixel radius, and each group of matching points in the center area constitutes the target matching point set. For example, reference Figure 3 As shown, the elliptical area A is composed of the center coordinates, the major axis pixel radius and the minor axis pixel radius, and the circular area B is composed of the center coordinates and the minor axis pixel radius. The matching points in the area (i.e., area C) of the elliptical area A that exceeds the circular area B are filtered out, and only the matching points in the circular area B are retained, and each group of matching points in the center area B constitutes a target matching point set.

[0055] Step S33, calculating the rotation angle sequence corresponding to each group of matching points in the target matching point set, and calculating the rotation angle of the current frame surgical image relative to the reference image based on the rotation angle sequence.

[0056] For the specific calculation of the rotation angle sequence corresponding to each group of matching points in the target matching point set, any method in the prior art can be used, for example, using vector cross product or rotation matrix method to calculate. Figure 4 As shown, the rotation angle of the current frame surgical image relative to the reference image is calculated based on the rotation angle sequence, including the following steps 331 to 332.

[0057] Step S331, calculating the median in the rotation angle sequence, and setting a threshold, eliminating the rotation angle sequence whose difference with the median is greater than the threshold, to obtain a corrected rotation angle sequence.

[0058] Step S332: taking the median of the corrected rotation angle sequence or the mean of the corrected rotation angle sequence as the rotation angle.

[0059] For example, the rotation angle sequences corresponding to each group of matching points in the target point set include: b1, b2, b3, b4, b5, b6, b7. The rotation angle sequences are sorted from small to large as b3, b1, b2, b7, b4, b6, b5, and the median b7 in the rotation angle sequence is obtained. A threshold T is set, and the difference between each rotation angle sequence and the median b7 is calculated. If it is greater than the threshold T, it is eliminated. If it is less than or equal to the threshold T, it is retained. The obtained values ​​are b3, b1, b2, b7, and b4, which are the corrected rotation angle sequences. The median in the corrected rotation angle sequence (i.e., the rotation angle sequence b2) or the mean in the corrected rotation angle sequence (i.e., ) as the rotation angle.

[0060] After the rotation angle is obtained, the target incision angle or the target astigmatism axis implantation angle is calculated based on the rotation angle and the reference navigation parameter. Target astigmatism axis implant angle It is the rotation angle.

[0061] After testing, an iris segmentation data set was constructed based on the collected surgical images to construct a registration test set. The registration was performed through a large registration model, and the rotation angle of the image in the registration test set relative to the reference image was manually screened and determined. The rotation angle was used as the gold standard. The present application performed registration on the same registration test set to obtain the predicted rotation angle. The absolute value of the angle difference between the predicted rotation angle and the gold standard was calculated, and the average error was calculated on all test sets. It was found that the average error was less than 1.5°. This shows that the accuracy of the rotation angle calculation in the ophthalmic surgery navigation method disclosed in the present application ensures the accuracy of the target incision angle and the target astigmatism axis implantation angle calculation.

[0062] Step S4: If the current frame of the surgical image is in the capsulorhexis stage, the pixel radius of the target capsulorhexis circle is calculated based on the iris fitting result and the reference navigation parameters.

[0063] Among them, the pixel radius of the target capsulorhexis circle WTW refers to the physical diameter of the irises of both eyes of the patient.

[0064] In summary, in the present application, the surgical images and biological parameters of both eyes of the patient are first obtained, and the reference image is determined. The reference navigation parameters corresponding to the reference image are configured according to the biological parameters. The eyeball iris segmentation model is constructed based on the surgical image. The current frame surgical image and the reference image in the surgical image are input into the eyeball iris segmentation model, and the iris segmentation results corresponding to the current frame surgical image and the reference image are output respectively. Different processing is performed according to the different stages in which the current frame surgical image is located. Specifically, if the current frame surgical image is in the incision stage, the iris segmentation result is fitted with an iris ellipse to obtain an iris fitting result, and the image registration calculation is performed on the current frame surgical image and the reference image according to the iris fitting result to obtain the rotation angle of the current frame surgical image relative to the reference image, and the target incision angle is calculated based on the rotation angle and the reference navigation parameters. If the current frame surgical image is in the capsulorhexis stage, the iris segmentation result is fitted with an iris circle to obtain an iris fitting result, and the pixel radius of the target capsulorhexis circle is calculated based on the iris fitting result and the reference image parameters. If the current frame of the surgical image is in the astigmatism axis stage, the iris segmentation result is fitted with an iris ellipse to obtain an iris fitting result, and the current frame of the surgical image and the reference image are image registered based on the iris fitting result to obtain the rotation angle of the current frame of the surgical image relative to the reference image, and the target astigmatism axis implantation angle is calculated based on the rotation angle and the reference navigation parameter. Therefore, in the incision stage or the astigmatism axis implantation stage, the current frame of the surgical image and the reference image are image registered based on the iris fitting result. Compared with the image registration method based on the iris blood vessels in the prior art, it is not limited by the local area information of the iris blood vessels, and at the same time, the accuracy of the target navigation parameters finally obtained is guaranteed, thereby solving the problem that the existing ophthalmic surgical navigation method is difficult to effectively solve the complex interference situations such as instrument occlusion during surgery and iris blood vessel deformation caused by surgical operation, and cannot guarantee the stability and accuracy of navigation.

[0065] In one embodiment, the ophthalmic surgery navigation method further includes: drawing the current frame surgical image based on the target incision angle or the target astigmatism axis implantation angle or the target capsulorhexis circle pixel radius, and performing 3D interlaced rendering on the drawn current frame surgical image, and inputting the 3D interlaced rendered current frame surgical image into a display device. Through 3D interlaced rendering and inputting the interlaced rendered current frame surgical image into a display device, a clinician wears a display device to achieve real-time observation of the target navigation parameters to ensure the real-time and high-instruction of the observation, thereby solving the problem that the existing 2D observation cannot meet the actual use needs of clinicians, and the 3D navigation observation based on point cloud information processing cannot take into account both real-time and high quality. Among them, the display device can be an ultra-high-definition polarized 3D medical monitor or a naked-eye 3D display device, which is not specifically limited in this embodiment. When drawing the current frame surgical image in the capsulorhexis stage, the capsulorhexis circle is drawn on the current frame surgical image according to the center coordinates obtained by iris circle fitting and the calculated target capsulorhexis circle pixel radius.

[0066] Specifically, since the surgical image is in a 3D left-right format or a 3D top-bottom format, the current frame surgical image includes: a first current frame surgical image and a second current frame surgical image, and accordingly, the reference image includes: a first reference image and a second reference image. The current frame surgical image can be drawn based on the target incision angle or the target astigmatism axis implantation angle or the target capsulorhexis pixel radius, and two methods can be used for drawing. One method is to calculate the target navigation parameters corresponding to the first current frame surgical image through the first current frame surgical image and the first reference image, and draw the first current frame surgical image and the second current frame surgical image based on the target navigation parameters. The other method is to calculate the target navigation parameters corresponding to the first current frame surgical image and the second current frame surgical image respectively through the first current frame surgical image and the first reference image, the second current frame surgical image and the second reference image, and draw the first current frame surgical image and the second current frame surgical image respectively based on the target navigation parameters corresponding to the first current frame surgical image and the second current frame surgical image respectively. Among them, the target navigation parameters include: the target incision angle or the target astigmatism axis implantation angle or the target capsulorhexis circle pixel radius.

[0067] The series of detailed descriptions listed above are only specific descriptions of feasible implementation methods of the present invention. They are not intended to limit the scope of protection of the present invention. Any equivalent implementation methods or changes that do not deviate from the technical spirit of the present invention should be included in the scope of protection of the present invention.

[0068] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

[0069] In addition, it should be understood that although the present specification is described according to implementation modes, not every implementation mode contains only one independent technical solution. This description of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment may also be appropriately combined to form other implementation modes that can be understood by those skilled in the art.

Claims

1. An ophthalmic surgery navigation method, characterized in that: include: Obtain surgical images and biological parameters of both eyes of the patient, determine a reference image, configure reference navigation parameters corresponding to the reference image according to the biological parameters, and construct an eyeball iris segmentation model based on the surgical images; Inputting the current frame surgical image and the reference image in the surgical image into the eyeball iris segmentation model, outputting iris segmentation results corresponding to the current frame surgical image and the reference image respectively, and performing iris fitting on the iris segmentation results to obtain iris fitting results; If the current frame of the surgical image is in the incision stage or the astigmatism axis implantation stage, perform image registration calculation on the current frame of the surgical image and the reference image according to the iris fitting result to obtain a rotation angle of the current frame of the surgical image relative to the reference image, and calculate a target incision angle or a target astigmatism axis implantation angle based on the rotation angle and the reference navigation parameters; If the current frame of the surgical image is in the capsulorhexis stage, the pixel radius of the target capsulorhexis circle is calculated based on the iris fitting result and the reference navigation parameters.

2. The ophthalmic surgery navigation method according to claim 1, characterized in that: The reference image is defined by a surgical image or a preoperative examination image captured from the surgical image.

3. The ophthalmic surgery navigation method according to claim 1, characterized in that: The performing iris fitting on the iris segmentation result to obtain an iris fitting result includes: If the current frame surgical image is in the incision stage or the astigmatism axis implantation stage, an iris ellipse fitting is performed on the iris segmentation results corresponding to the current frame surgical image and the reference image respectively, to obtain an iris fitting result consisting of the center coordinates, the major axis pixel radius and the minor axis pixel radius; If the current frame surgical image is in the capsulorhexis stage, iris circle fitting is performed on the iris segmentation results corresponding to the current frame surgical image and the reference image respectively to obtain an iris fitting result consisting of circle center coordinates and pixel radius.

4. The ophthalmic surgery navigation method according to claim 3, characterized in that: The performing image registration calculation on the current frame surgical image and the reference image according to the iris fitting result to obtain the rotation angle of the current frame surgical image relative to the reference image includes: Based on the iris fitting result, the current frame surgical image and the reference image are respectively cropped to obtain an image to be registered and a registration reference image; Performing a fusion calculation on the image to be registered and the registration reference image to obtain a matching point set consisting of multiple groups of matching points, and screening the matching points in the matching point set according to the iris fitting result to obtain a target matching point set; The rotation angle sequence corresponding to each group of matching points in the target matching point set is calculated, and the rotation angle of the current frame surgical image relative to the reference image is calculated based on the rotation angle sequence.

5. The ophthalmic surgery navigation method according to claim 4, characterized in that: The method of cropping the current frame surgical image and the reference image based on the iris fitting result to obtain the image to be registered and the registered reference image includes: With the center coordinates of the circle as the center point, crop the current frame surgical image and the reference image with a width and height of 2r * ×k(r * ∈{r l ,r s }) as the image to be registered and the registration reference image respectively; Where k is the cropping ratio, and k>1, r l refers to the major axis pixel radius, r s refers to the minor axis pixel radius.

6. The ophthalmic surgery navigation method according to claim 4, characterized in that: The step of screening the matching points in the matching point set according to the iris fitting result to obtain a target matching point set includes: The matching points in the circular area formed by the coordinates of the circle center and the short axis pixel radius are retained, and each group of matching points in the circular area constitutes the target matching point set.

7. The ophthalmic surgery navigation method according to claim 4, characterized in that: The step of calculating the rotation angle of the current frame surgical image relative to the reference image based on the rotation angle sequence comprises: Calculating the median of the rotation angle sequence, setting a threshold, and eliminating the rotation angle sequence whose difference with the median is greater than the threshold, to obtain a corrected rotation angle sequence; The median of the corrected rotation angle sequence or the mean of the corrected rotation angle sequence is used as the rotation angle.

8. The ophthalmic surgery navigation method according to claim 1, characterized in that: The reference navigation parameters include: incision angle Capsulorhexis circle physical diameter R gt and implantation angle The target incision angle Target astigmatism axis implantation angle The target capsulorhexis circle pixel radius refers to the rotation angle, and WTW refers to the physical diameter of the irises of both eyes of the patient.

9. The ophthalmic surgery navigation method according to claim 1, characterized in that: The ophthalmic surgery navigation method also includes: drawing the current frame surgical image based on the target incision angle or the target astigmatism axis implantation angle or the target capsulorhexis circle pixel radius, and performing 3D interlaced rendering on the drawn current frame surgical image, and inputting the 3D interlaced rendered current frame surgical image into a display device.

10. The ophthalmic surgery navigation method according to claim 9, characterized in that: The surgical image is in a 3D left-right format or a 3D top-bottom format, and the current frame surgical image includes: a first current frame surgical image and a second current frame surgical image; Drawing the current frame surgical image based on the target incision angle or the target astigmatism axis implantation angle or the target capsulorhexis circle pixel radius includes: Calculating target navigation parameters corresponding to the first current frame surgical image, and drawing the first current frame surgical image and the second current frame surgical image respectively based on the target navigation parameters; or calculating target navigation parameters corresponding to the first current frame surgical image and the second current frame surgical image respectively, and drawing the first current frame surgical image and the second current frame surgical image respectively based on the target navigation parameters corresponding to the first current frame surgical image and the second current frame surgical image respectively; Wherein, the target navigation parameters include: the target incision angle or the target astigmatism axis implantation angle or the target capsulorhexis circle pixel radius.

Images