Laser therapy equipment control method and treatment system based on optical imaging guidance

By calculating the possibility and neighborhood approximation of the points to be grown in the seed growth algorithm, the problem of insufficient vascular positioning accuracy in the prior art is solved, and the accuracy and safety of laser treatment are improved.

CN118806225BActive Publication Date: 2024-11-22CHANGCHUN ZHONGKE CHANGGUANG SPATIOTEMPORAL PHOTOELECTRIC TECH CO LTD +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411299002.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-18
Publication Date
2024-11-22
Estimated Expiration
2044-09-18

AI Technical Summary

Technical Problem

When using seed growth algorithm to locate the fundus images, the prior art cannot accurately process the characteristics of the vascular color gradually becoming lighter along the end direction and the color depth of non-vascular positions, resulting in a decrease in the accuracy of laser treatment, which may cause healthy blood vessels. damage.

Method used

In the seed growth algorithm, by calculating the possibility of the point to be grown, only points with a possibility greater than the preset threshold are grown, and the approximation between the neighboring pixel points of the point to be grown and the pixel points in the growth path are taken into account. The deviation between approximation of the approximation of the approximation between the pixel points in the neighborhood of the points to be grown is measured using gradient direction consistency, and the accuracy of the vascular pixel points is improved.

Benefits of technology

The accuracy of vascular pixel points in fundus images processed by seed growth algorithm is improved, the accuracy of fundus blood vessel position is enhanced, unnecessary damage to healthy tissue is avoided, and the effectiveness and safety of treatment are improved.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118806225B_ABST
    Figure CN118806225B_ABST
Patent Text Reader

Abstract

The present invention relates to the field of image processing technology. Specifically, it relates to a laser treatment device control method and treatment system based on optical imaging guidance. The method includes: collecting fundus images, determining the positions of blood vessel pixels in the fundus images using a seed growth algorithm, and then adjusting the parameters of the laser treatment device using the positions of blood vessel pixels, wherein the positions of blood vessel pixels in the fundus images using the seed growth algorithm include: obtaining an approximate blood vessel connected domain; calculating the selection value of each approximate blood vessel connected domain, and taking the pixel point with the smallest grayscale value in the approximate blood vessel connected domain whose selection value is greater than a preset selection value threshold as a seed point, and taking the pixel point in the neighborhood of the seed point as a point to be grown; calculating the possibility that the point to be grown is a blood vessel pixel point, and growing the point to be grown whose possibility is greater than the preset possibility threshold. The method of the present invention can improve the accuracy of positioning fundus blood vessels.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention generally relates to the field of image processing technology. More specifically, the present invention relates to a laser treatment device control method and treatment system based on optical imaging guidance. Background Art

[0002] When treating dry eye, doctors can accurately determine the lesion area and treatment point of the dry eye patient's eye through real-time optical imaging guidance, so as to accurately irradiate the laser to the target tissue and maximize the accuracy and effectiveness of the treatment. Based on the acquired eye images and the patient's clinical condition, doctors can develop personalized treatment plans and adjust the parameters and duration of laser treatment according to the patient's condition and treatment goals, so as to maximize the treatment effect and reduce unnecessary damage.

[0003] When treating patients with dry eyes, optical imaging equipment is usually used first to obtain images of the eye structure in order to observe the reactions and changes of the eye tissues during treatment, helping doctors to adjust treatment strategies in a timely manner and ensure the safety and effectiveness of the treatment. Through precise optical imaging guidance and monitoring, possible complications during laser treatment, such as corneal damage and infection, can be reduced, improving the safety of treatment. It can provide patients with a more comfortable and safe treatment experience, reduce patients' anxiety and discomfort, and improve treatment acceptance and satisfaction. Laser treatment methods for dry eyes guided by optical imaging can play an important role in the treatment of dry eyes, improve the accuracy, personalization and safety of treatment, provide patients with a more effective and comfortable treatment experience, and achieve better treatment results.

[0004] The patient's eye structure image is monitored in real time through an optical imaging device, and then the blood vessels in the eye structure image are segmented. Then, the parameters of the laser treatment device are adjusted according to the segmented blood vessels, and the patient's eyes are treated. This real-time monitoring and adjustment feature can ensure the accuracy and safety of the treatment process, effectively avoiding damage to healthy tissues; it fully considers the individual differences of patients and the diversity of eye lesions. The control system adjusts the treatment parameters of the laser treatment device according to the patient's eye structure image data, achieving targeted personalized treatment, improving the effect and applicability of the treatment, and achieving precise positioning of the treatment. This means that laser treatment can accurately target the lesion area of ​​the patient's eye, avoiding unnecessary damage to healthy tissues, thereby improving the effect and safety of the treatment.

[0005] Because the color of the end of the blood vessels on the eyeball is extremely similar to the fundus and the blood vessels have a color gradient, the blood vessels in the eye structure image captured by the optical imaging device are relatively blurred. In order to increase the clarity of the blood vessels in the image and better locate the blood vessels, the prior art usually uses a seed growth algorithm to process the eye structure image after acquiring it to make the blood vessels in the image clearer. However, due to the characteristics of the blood vessel color in the fundus image gradually becoming lighter along the end of the blood vessel and the color of the pixels at the non-vascular position of the fundus being different in depth and having a gradient, and the non-fundus position near the end of the blood vessel is extremely similar in color to the pixel points at the end of the blood vessel, the traditional seed growth algorithm cannot accurately grow the pixel points at the end of the blood vessel, resulting in a decrease in the accuracy of laser treatment for dry eye syndrome. Because the accurate positioning and segmentation of the treatment cannot be achieved, it may cause damage to healthy blood vessels, reducing the safety of the treatment.

[0006] For example, a Chinese patent with authorization announcement number CN109410191B discloses a fundus vascular positioning method based on OCT images, which uses a seed growing algorithm to segment the blood vessels in the fundus image. Although this method can locate the blood vessels, when processing the fundus image using the seed growing algorithm, this method does not take into account the characteristics of the blood vessel color in the fundus image gradually becoming lighter along the end of the blood vessel, the color of the pixel points at the non-vascular position of the fundus is different in depth and there is a gradient, and the color of the non-fundus position near the end of the blood vessel is extremely similar to the color of the pixel points at the end of the blood vessel. Therefore, the pixel points at the end of the blood vessel cannot be accurately grown, and there is still a certain error in the positioning of the blood vessel. Summary of the invention

[0007] To solve one or more of the above technical problems, the present invention provides solutions in the following aspects.

[0008] In a first aspect, the present invention provides a method for controlling a laser treatment device based on optical imaging guidance, comprising:

[0009] Collect fundus images, use a seed growth algorithm to determine the positions of blood vessel pixels in the fundus images, and then use the positions of blood vessel pixels to adjust the parameters of the laser treatment device, wherein using the seed growth algorithm to determine the positions of blood vessel pixels in the fundus images includes:

[0010] Connected area detection is performed on the fundus image to obtain an approximate vascular connected domain; a selection value of each approximate vascular connected domain is calculated, and a pixel point with the smallest grayscale value in the approximate vascular connected domain whose selection value is greater than a preset selection value threshold is used as a seed point, and pixels in the neighborhood of the seed point are used as points to be grown; the selection value is negatively correlated with the grayscale mean of the edge pixels of the corresponding connected domain and the gradient mean of its internal pixels;

[0011] The possibility that the point to be grown is a blood vessel pixel is calculated based on the deviation between the approximation between the neighboring pixel points of the point to be grown and the approximation between the neighboring pixel points of the pixel points in its growth path, and the possibility is negatively correlated with the deviation; the approximation between the neighboring pixel points of a certain pixel point is positively correlated with the consistency of the gradient direction of the pixel point and its related pixel points, and the related pixel points are located in the gradient direction of the pixel point and the distance from the pixel point is less than a preset threshold;

[0012] The points to be grown whose likelihood is greater than a preset likelihood threshold are grown.

[0013] The beneficial effect is as follows: before using the seed growth algorithm to grow the growth points, the control method of the laser treatment equipment based on optical imaging guidance in this embodiment first calculates the possibility that each pixel point to be grown is a blood vessel pixel point, and only grows the growth points whose possibility is greater than a preset possibility threshold, thereby avoiding to a certain extent the misjudgment of pixels that do not belong to blood vessels as blood vessel pixels. In addition, when calculating the possibility that the pixel point to be grown is a blood vessel pixel point, the relationship between the deviation between the approximation between the neighboring pixels of the growth point and the approximation between the neighboring pixels of the pixel points in its growth path and the possibility that the pixel point to be grown is a blood vessel pixel point is taken into account, thereby making the calculation result of the possibility that the pixel point to be grown is a blood vessel pixel point more accurate, thereby improving the accuracy of the blood vessel pixels in the fundus image processed by the seed growth algorithm, and then improving the accuracy of the obtained fundus blood vessel position, avoiding unnecessary damage to healthy tissue, thereby improving the effect and safety of treatment, and achieving better treatment effect. In addition, when calculating the approximation between the neighboring pixels of the point to be grown, the difference between the distribution of the gradient direction of the pixel points corresponding to the non-vascular fundus position and the distribution of the gradient direction of the pixel points corresponding to the vascular position is taken into account. The approximation between the neighboring pixels of the point to be grown is measured based on the consistency of the gradient direction of the point to be grown and its related pixels, so that the calculated approximation between the neighboring pixels of the point to be grown is more accurate, thereby further improving the accuracy of the vascular pixels in the fundus image processed by the seed growth algorithm.

[0014] In one embodiment, performing connected area detection on the fundus image includes:

[0015] Gray-scaling the fundus image to obtain a fundus grayscale image;

[0016] The collected fundus grayscale image is threshold segmented with a preset grayscale threshold to obtain a binary image, and the connected domain in the binary image is used as the approximate vascular connected domain.

[0017] In one embodiment, the possibility threshold is 0.5, and the selection value threshold is 0.8.

[0018] In one embodiment, the calculation expression for the possibility that the to-be-grown point is a blood vessel pixel is:

[0019] ;

[0020] In the formula, Indicates the possibility that the jth growth point is a blood vessel pixel. represents the cosine of the angle between the line connecting the jth growth point to be grown and the previous growth point, and the line connecting the jth growth point to be grown and the pth pixel point in its growth path; represents the approximation between the neighboring pixels of the jth point to be grown, Indicates the approximation between the neighboring pixels of the pth pixel in the growth path of the jth point to be grown, is the maximum and minimum normalization function.

[0021] The beneficial effect is as follows: in the present embodiment, when calculating the possibility that the point to be grown is a blood vessel pixel point, not only is the possibility of the point to be grown being a blood vessel pixel point measured by using the deviation between the approximation between the neighboring pixels of the point to be grown and the approximation between the neighboring pixels of the pixels in its growth path, but the relationship between the directionality of the growth path and the possibility that the point to be grown is a blood vessel pixel point is further considered, thereby improving the accuracy of the calculation result of the possibility that the point to be grown is a blood vessel pixel point.

[0022] In one embodiment, calculating the similarity between neighboring pixels of a point to be grown includes:

[0023] The pixel points in the gradient direction of the point to be grown in the window centered on the point to be grown are taken as the relevant pixel points of the point to be grown, and the number of pixel points in the window located in the gradient direction of the point to be grown and having a grayscale difference with the point to be grown that is less than a preset grayscale difference value is recorded as A;

[0024] The gradient direction consistency of the point to be grown and its related pixel points is recorded as B, and the approximation of the point to be grown and its related pixel points is calculated based on B and A, and recorded as E. Then, the approximation between the neighborhood pixel points of the point to be grown is calculated based on E. The approximation between the neighborhood pixel points of the point to be grown is positively correlated with E, and E is positively correlated with both B and A.

[0025] The beneficial effect is that the grayscale of the pixels inside the blood vessel position is very similar, while the grayscale of the pixels at the non-blood vessel fundus position varies more. In this embodiment, when calculating the approximation of the pixels in the neighborhood of the point to be grown, not only the gradient direction consistency of the point to be grown and its related pixels is taken into account, but also the number of pixels in the window centered on the point to be grown whose grayscale difference with the point to be grown is less than the preset grayscale difference value. Since the more the number of the above-mentioned pixels is, the more similar the grayscales of different pixels in the window are, therefore, by making the approximation of the point to be grown and its related pixels positively correlated with the number of pixels of the first approximate grayscale level, the accuracy of the approximation calculation result of the pixels in the neighborhood of the point to be grown is further improved.

[0026] In one embodiment, the grayscale difference preset value is 5.

[0027] In one embodiment, the calculation expression of the approximation between the to-be-grown point and its related pixel point is:

[0028] ;

[0029] In the formula, Indicates the approximation between the jth point to be grown and its related pixels, Indicates the number of related pixels of the jth growth point j, The unit vector representing the gradient direction of the nth related pixel of the jth point to be grown, Represents the unit vector of the gradient direction of the jth growth point, Represents the number of first approximate grayscale pixels, Represents the grayscale difference between the jth point to be grown and its nth related pixel.

[0030] Its beneficial effect is that when calculating the approximation between the point to be grown and its related pixel points, the difference between the grayscale of the point to be grown and its related pixel points is further taken into account, and the greater the difference between the grayscales of the two, the greater the similarity between the two, thereby improving the accuracy of the calculated approximation between the point to be grown and its related pixel points, and then improving the accuracy of the calculation result of the approximation between the neighborhood pixel points of the point to be grown.

[0031] In one embodiment, the calculating the similarity between neighboring pixels of a point to be grown further includes:

[0032] The pixel points in the gradient direction of the previous growth point of the growth point to be grown in the window are recorded as the relevant pixel points of the previous growth point; the number of pixel points whose grayscale difference with the previous growth point is less than the preset grayscale difference value is recorded as C;

[0033] The gradient direction consistency between the previous growth point and its related pixel points is recorded as D. The approximation between the previous growth point and its related pixel points is calculated based on C and D and recorded as F. The product of E and F is used as the approximation between the neighborhood pixel points of the growth point to be grown; F is positively correlated with both C and D.

[0034] The beneficial effect is that the product of the approximation between the pixel points in the window to be grown and its gradient direction and the approximation between the pixel points in the previous growth point and its gradient direction can more comprehensively reflect the similarity between all the pixel points in the window corresponding to the point to be grown. Therefore, the product of the two approximations is used as the approximation between the neighborhood pixel points of the point to be grown, so that the calculation result of the approximation between the neighborhood pixel points of the point to be grown is more accurate.

[0035] In one embodiment, the calculation expression of the selected value is:

[0036] ;

[0037] In the formula, represents the selected value of the i-th approximate vascular connectivity domain, represents the maximum distance between pixels in the i-th approximate vascular connected domain, represents the grayscale mean of the edge pixels of the i-th approximate vascular connected domain, Represents the gradient mean of the internal pixels of the i-th approximate vascular connected domain.

[0038] The beneficial effect is that since the blood vessels in the fundus are usually long, the calculation expression of the selection value of the present invention not only takes into account the grayscale mean of the edge pixels of the approximate blood vessel connected domain and the gradient mean of its internal pixels when calculating the selection value, but also further takes into account the length of the approximate blood vessel connected domain, thereby making the calculation result of the selection value more accurate.

[0039] In a second aspect, the present invention provides a laser treatment system, comprising an optical imaging device, a laser treatment device and a control system, wherein the optical imaging device is used to acquire fundus images, the laser treatment device is used to apply laser treatment to the eye, and the control system is used to execute the laser treatment device control method based on optical imaging guidance of the present invention.

[0040] In summary, the beneficial effects of the present invention are as follows: the method of the present invention can greatly improve the accuracy of the blood vessel pixels in the fundus image processed by the seed growth algorithm, thereby improving the accuracy of the obtained fundus blood vessel position. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein:

[0042] Figure 1 is a flow chart schematically showing a method for controlling a laser treatment device based on optical imaging guidance according to an embodiment of the present invention;

[0043] Figure 2 is a schematic diagram of a fundus grayscale image schematically illustrating an embodiment of the present invention;

[0044] Figure 3 is a schematic diagram schematically showing a flow chart of a method for calculating the approximation between neighboring pixels of a point to be grown according to an embodiment of the present invention;

[0045] Figure 4 is a schematic diagram schematically showing the structure of a laser treatment system according to an embodiment of the present invention;

[0046] Figure 5 It is a schematic diagram schematically showing the structure of a control system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0047] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0048] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0049] Embodiment of the laser treatment device control method based on optical imaging guidance:

[0050] like Figure 1 As shown, the laser treatment device control method based on optical imaging guidance of the present invention includes: collecting fundus images, using a seed growth algorithm to determine the positions of blood vessel pixels in the fundus images, and then using the positions of the blood vessel pixels to adjust the parameters of the laser treatment device, wherein using the seed growth algorithm to determine the positions of blood vessel pixels in the fundus images includes:

[0051] S101, determining seed points and points to be grown, specifically: performing connected area detection on the fundus image to obtain an approximate vascular connected domain; calculating a selection value of each approximate vascular connected domain, and taking a pixel point with the smallest grayscale value in the approximate vascular connected domain whose selection value is greater than a preset selection value threshold as a seed point, and taking pixels in the neighborhood of the seed point as points to be grown; the selection value is negatively correlated with the grayscale mean of the edge pixels of the corresponding connected domain and the gradient mean of its internal pixels.

[0052] Fundus images are images of the eye collected using professional medical instruments.

[0053] Because the grayscale value of the pixel point at the intraocular blood vessel position is low, and compared with the non-vascular fundus position, the intraocular blood vessel has a large grayscale change only at the edge position and a small internal grayscale change, while the grayscale of the pixel point at the non-vascular fundus position changes faster and with a larger amplitude. In addition, the grayscale value of the pixel point at the internal position of the intraocular blood vessel is low and very similar, and the gradient of the pixel point corresponding to the internal position of the blood vessel is small. Therefore, making the selected value negatively correlated with the grayscale mean of the edge pixel points of the corresponding approximate blood vessel connected domain and the gradient mean of its internal pixel points can more accurately calculate the selected value of each approximate blood vessel connected domain.

[0054] S102, calculating the possibility that the point to be grown is a blood vessel pixel, specifically: calculating the possibility that the point to be grown is a blood vessel pixel according to the deviation between the approximation between the neighboring pixel points of the point to be grown and the approximation between the neighboring pixel points of the pixel points in its growth path, wherein the possibility is negatively correlated with the deviation; the approximation between the neighboring pixel points of a certain pixel point is positively correlated with the consistency of the gradient direction of the pixel point and its related pixel points, and the related pixel point is located in the gradient direction of the pixel point and the distance from the pixel point is less than a preset threshold;

[0055] Normally, there is a large difference in feature quantities such as grayscale and gradient between pixel points at non-vascular fundus positions, while the characteristics of pixel points at vascular positions are highly similar, so the features of pixel points at each vascular position during the growth process will also be very similar; if the deviation between the similarity between the neighborhood pixel points of the point to be grown and the similarity between the neighborhood pixel points of the pixel points in its growth path is large, it means that the point to be grown is likely to be a pixel point far away from the inside of the blood vessel, and the probability that the point to be grown is a pixel point in the blood vessel is lower. Therefore, the possibility that the point to be grown is a vascular pixel can be calculated based on the deviation between the similarity between the neighborhood pixel points of the point to be grown and the similarity between the neighborhood pixel points of the pixel points in its growth path, and the possibility is negatively correlated with the deviation.

[0056] In addition, the gradient direction of the pixel points at the non-vascular fundus position is relatively dispersed. However, the overall grayscale of the pixel points inside the vascular position is very similar, and the color change gradually becomes lighter along the end direction of the blood vessel, so the gradient direction of the pixel points inside the blood vessel will be relatively similar along the end direction of the blood vessel. Therefore, the consistency of the gradient direction of the point to be grown and its related pixels can be used to measure the similarity between the neighboring pixels of the point to be grown, and then measure the degree to which the point to be grown is close to the inside of the blood vessel.

[0057] S103, growing the points to be grown whose likelihood is greater than a preset likelihood threshold.

[0058] After growing the points to be grown whose probability is greater than the preset probability threshold, the pixel points in the pixel neighborhood range are new points to be grown, and then step S102 and step S103 are performed again to obtain new points to be grown again, and step S102 and step S103 are repeatedly performed iteratively until the probability of all points to be grown is less than the preset probability threshold. Finally, the positions of all blood vessel pixels in the fundus image can be obtained.

[0059] The explanation of the growth path in this embodiment is as follows: assuming that the current growth is the third growth, pixel C is the point to be grown, C is a pixel in the neighborhood of pixel B participating in the growth for the second time, and pixel B is a pixel in the neighborhood of pixel A participating in the growth for the first time, and pixel A is a pixel in the neighborhood of the seed point. The corresponding growth path is seed point-pixel A-pixel B-pixel C.

[0060] In this embodiment, the preset possibility threshold is 0.5, and in other embodiments, the preset possibility threshold may also be other appropriate values. In this embodiment, the preset selection value threshold is 0.8, and in other embodiments, the preset selection value threshold may also be other appropriate values.

[0061] The control method of the laser treatment equipment based on optical imaging guidance in this embodiment first calculates the possibility that each pixel point to be grown is a blood vessel pixel point before using the seed growth algorithm to grow the growth point, and only grows the growth point whose possibility is greater than a preset possibility threshold, thereby avoiding to a certain extent the misjudgment of pixels that do not belong to blood vessels as blood vessel pixels. In addition, when calculating the possibility that the pixel point to be grown is a blood vessel pixel point, the relationship between the deviation between the approximation between the neighboring pixels of the growth point and the approximation between the neighboring pixels of the pixel points in its growth path and the possibility that the pixel point to be grown is a blood vessel pixel point is taken into account, thereby making the calculation result of the possibility that the pixel point to be grown is a blood vessel pixel point more accurate, thereby improving the accuracy of the blood vessel pixels in the fundus image processed by the seed growth algorithm, and then improving the accuracy of the obtained fundus blood vessel position, avoiding unnecessary damage to healthy tissue, thereby improving the effect and safety of treatment, and achieving better treatment effect. In addition, when calculating the approximation between the neighboring pixels of the point to be grown, the difference between the distribution of the gradient direction of the pixel points corresponding to the non-vascular fundus position and the distribution of the gradient direction of the pixel points corresponding to the vascular position is taken into account. The approximation between the neighboring pixels of the point to be grown is measured based on the consistency of the gradient direction of the point to be grown and its related pixels, so that the calculated approximation between the neighboring pixels of the point to be grown is more accurate, thereby further improving the accuracy of the vascular pixels in the fundus image processed by the seed growth algorithm.

[0062] In one embodiment, performing connected area detection on the fundus image includes:

[0063] S201 , grayscale the fundus image to obtain a fundus grayscale image.

[0064] The obtained fundus grayscale image is as follows Figure 2 shown.

[0065] S202, performing threshold segmentation on the collected fundus grayscale image with a preset grayscale threshold to obtain a binary image and using the connected domain in the binary image as an approximate vascular connected domain.

[0066] Since the grayscale values ​​of pixels at the vascular position in the fundus grayscale image are lower, and the grayscale values ​​of pixels at the non-vascular position are higher, the connected domain with darker color obtained by threshold segmentation of the fundus grayscale image is likely to be the vascular area.

[0067] It can be seen from the above embodiments that the possibility of the point to be grown being a blood vessel pixel can be calculated based on the deviation between the approximation between the neighboring pixels of the point to be grown and the approximation between the neighboring pixels of the pixel points in its growth path, and the possibility of the point to be grown being a blood vessel pixel is negatively correlated with the deviation; in one embodiment, the calculation expression of the possibility of the point to be grown being a blood vessel pixel is:

[0068]

[0069] In the formula, Indicates the possibility that the jth growth point is a blood vessel pixel. represents the cosine of the angle between the line connecting the jth growth point to be grown and the previous growth point, and the line connecting the jth growth point to be grown and the pth pixel point in its growth path; represents the approximation between the neighboring pixels of the jth point to be grown, Indicates the approximation between the neighboring pixels of the pth pixel in the growth path of the jth point to be grown, is the maximum and minimum normalization function.

[0070] The larger the value is, the smaller the angle between the line connecting the jth growth point and the previous growth point, and the line connecting the jth growth point and the pth pixel point in its growth path. This means that the directionality of the growth path is stronger (i.e., the growth path is closer to a straight line). Because the characteristics of the pixel points at the blood vessel position have a high approximation, the gradient of the pixel points at the blood vessel position has a strong directionality, so The larger the value is, the greater the possibility that the jth growth point is a blood vessel pixel.

[0071] Generally, when using the seed growth algorithm to process an image, it is necessary to first determine the seed point, start growing from the seed point to its neighborhood, obtain the growth point, and then further grow to the neighborhood of the growth point. In this embodiment, the growth point whose growth moment is closest to the current moment is recorded as the previous growth point, and the pixel point in the neighborhood of the previous growth point is the point to be grown. If the growth at the current moment is the first growth iteration process, the corresponding seed point is used as the previous growth point. In this embodiment, when calculating the possibility that the point to be grown is a blood vessel pixel, not only is the deviation between the approximation between the neighboring pixel points of the point to be grown and the approximation between the neighboring pixel points of the pixel points in its growth path considered to measure the possibility that the point to be grown is a blood vessel pixel, but also the relationship between the directionality of the growth path and the possibility that the point to be grown is a blood vessel pixel is further considered, thereby improving the accuracy of the calculation result of the possibility that the point to be grown is a blood vessel pixel.

[0072] It can be seen from the above embodiments that the similarity between the neighboring pixels of the point to be grown is positively correlated with the consistency of the gradient direction of the point to be grown and its related pixels, such as Figure 3 As shown, in one embodiment, calculating the approximation between neighboring pixels of a point to be grown includes:

[0073] S301. Obtain the number of first approximate grayscale pixel points, specifically: take the pixel points in the gradient direction of the point to be grown in a window centered on the point to be grown as the relevant pixel points of the point to be grown, and record the number of pixel points in the window that are located in the gradient direction of the point to be grown and whose grayscale difference with the point to be grown is less than a preset grayscale difference value as A.

[0074] In this embodiment, the window size can be set to 5 5. In other embodiments, the size of the window can also be set to other suitable sizes. In this embodiment, the grayscale difference preset value can be 5. In other embodiments, the grayscale difference preset value can also be other suitable values. In this embodiment, the pixel points in the window are used as the neighborhood pixel points of the point to be grown.

[0075] S302, calculating the approximation between the neighboring pixels of the point to be grown, specifically: the gradient direction consistency of the point to be grown and its related pixels is recorded as B, the approximation of the point to be grown and its related pixels is calculated based on B and A, and recorded as E, and then the approximation between the neighboring pixels of the point to be grown is calculated based on E, the approximation between the neighboring pixels of the point to be grown is positively correlated with E, and E is positively correlated with both B and A.

[0076] The consistency of the gradient direction of the to-be-grown point and its related pixel points refers to the consistency between the gradient direction of the to-be-grown point and the gradient direction of its related pixel points.

[0077] Since the grayscales of the pixels inside the blood vessel position are very similar, while the grayscales of the pixels at the non-blood vessel fundus position vary greatly. In this embodiment, when calculating the approximation of the pixels within the neighborhood of the point to be grown, not only the gradient direction consistency of the point to be grown and its related pixels is taken into account, but also the number of pixels within the window centered on the point to be grown whose grayscale difference with the point to be grown is less than the preset grayscale difference value is further taken into account. Since the more the number of the above-mentioned pixels is, the more similar the grayscales of different pixels within the window are, therefore, by making the approximation of the point to be grown and its related pixels positively correlated with the number of pixels of the first approximate grayscale level, the accuracy of the approximation calculation result of the pixels within the neighborhood of the point to be grown is further improved.

[0078] In one embodiment, the calculation expression of the approximation between the point to be grown and its related pixel point is:

[0079]

[0080] In the formula, Indicates the approximation between the jth point to be grown and its related pixels, Indicates the number of related pixels of the jth growth point j, The unit vector representing the gradient direction of the nth related pixel of the jth point to be grown, Represents the unit vector of the gradient direction of the jth growth point, Represents the number of first approximate grayscale pixels, Represents the grayscale difference between the jth point to be grown and its nth related pixel.

[0081] The calculation expression in this embodiment further takes into account the difference in grayscale between the point to be grown and its related pixels when calculating the approximation between the point to be grown and its related pixels, and the greater the difference in grayscale between the two, the greater the similarity between the two, thereby improving the accuracy of the calculated approximation between the point to be grown and its related pixels, and further improving the accuracy of the calculation result of the approximation of the neighborhood pixels of the point to be grown.

[0082] In one embodiment, calculating the similarity between neighboring pixels of a point to be grown further includes:

[0083] S401, record the pixel points in the gradient direction of the previous growth point of the to-be-grown point in the window as the relevant pixel points of the previous growth point; record the number of pixel points whose grayscale difference with the previous growth point is less than a preset grayscale difference value as C.

[0084] S402, the gradient direction consistency of the previous growth point and its related pixel points is recorded as D, the approximation of the previous growth point and its related pixel points is calculated based on C and D, and recorded as F, and the product of E and F is used as the approximation between the neighborhood pixel points of the growth point to be grown; F is positively correlated with C and D.

[0085] The calculation method of the approximation between the previous growth point and its related pixel point is the same as the calculation method of the approximation between the to-be-grown point and its related pixel point in the above embodiment, which will not be repeated here.

[0086] The product of the approximation between the point to be grown in the window and the pixel points in its gradient direction, and the approximation between the previous growth point and the pixel points in its gradient direction, can more comprehensively reflect the similarity between all the pixel points in the window corresponding to the point to be grown. Therefore, the product of the two approximations is used as the approximation between the neighborhood pixel points of the point to be grown, so that the calculation result of the approximation between the neighborhood pixel points of the point to be grown is more accurate.

[0087] It can be seen from the above embodiments that the approximation between the neighboring pixels of the point to be grown is positively correlated with the approximation between the point to be grown and its related pixels, and is positively correlated with the approximation between the previous growth point of the point to be grown and its related pixels. In one embodiment, the calculation expression for the approximation between the neighboring pixels of the point to be grown is:

[0088]

[0089] In the formula, Indicates the number of related pixels of the jth growth point j, The unit vector representing the gradient direction of the nth related pixel of the jth point to be grown, Represents the unit vector of the gradient direction of the jth growth point, Represents the number of first approximate grayscale pixels, It represents the grayscale difference between the jth growth point and its nth related pixel. Indicates the number of related pixels of the previous growth point of the jth growth point, Represents the number of pixels with the second approximate gray level. The unit vector representing the gradient direction of the nth related pixel point of the previous growth point of the jth growth point, The unit vector representing the gradient direction of the previous growth point of the jth growth point, Represents the grayscale difference between the previous growth point of the j-th growth point and its n-th related pixel.

[0090] It can be seen from the above embodiments that the selected value is negatively correlated with the grayscale mean of the edge pixels of the corresponding approximate blood vessel connected domain and the gradient mean of its internal pixels. In one embodiment, the calculation expression of the selected value is:

[0091]

[0092] In the formula, represents the selected value of the i-th approximate vascular connectivity domain, represents the maximum distance between pixels in the i-th approximate vascular connected domain, represents the grayscale mean of the edge pixels of the i-th approximate vascular connected domain, Represents the gradient mean of the internal pixels of the i-th approximate vascular connected domain.

[0093] The maximum pixel distance refers to the distance between the two pixels with the farthest distance in the connected domain.

[0094] The grayscale average of each edge pixel point of the i-th approximate vascular connected domain is obtained by averaging the grayscale average of the edge pixel points of the i-th approximate vascular connected domain, and the gradient average of each internal pixel point of the i-th approximate vascular connected domain is obtained by averaging the gradient of each internal pixel point of the i-th approximate vascular connected domain.

[0095] The size of the selection value can reflect the probability of the existence of vascular pixels in the approximate vascular connected domain. Since the blood vessels of the fundus are usually long, the calculation expression of the selection value in this embodiment not only takes into account the grayscale mean of the edge pixels of the approximate vascular connected domain and the gradient mean of its internal pixels when calculating the selection value, but also takes into account the length of the approximate vascular connected domain, so that the calculation result of the selection value is more accurate.

[0096] Laser therapy system example:

[0097] The present invention also provides a laser treatment system. Figure 4 As shown, the laser treatment system includes an optical imaging device, a laser treatment device and a control system. The optical imaging device is used to collect fundus images, the laser treatment device is used to apply laser treatment to the eye, and the control system is used to execute the laser treatment device control method based on optical imaging guidance as described in the first aspect of the present invention.

[0098] The optical imaging device monitors the patient's eye structure in real time and generates eye structure image data. The acquired eye structure image data is transmitted to the control system, which receives and uses the laser treatment device control method based on optical imaging guidance of the present invention to analyze the eye structure image data and automatically adjust the parameters of the laser treatment device, and the laser treatment device applies laser treatment to the eye. For example, according to the location, shape and depth of the lesion, the focal length, energy density and action time of the laser are adjusted to ensure that the laser can accurately act on the lesion area and achieve accurate positioning and adjustment of the treatment.

[0099] Optical imaging equipment can use optical coherence tomography (OCT), corneal topography or other imaging technologies to monitor the surface morphology and structure of the patient's eye in real time and generate eye structure image data.

[0100] like Figure 5 As shown, the control system includes a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the method for controlling the laser therapy device based on optical imaging guidance according to the first aspect of the present invention is implemented.

[0101] The control system also includes other components familiar to those skilled in the art, such as a communication bus and a communication interface, whose configuration and functions are known in the art and thus will not be described in detail here.

[0102] In the present invention, the aforementioned memory may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, apparatus or device. For example, a computer-readable storage medium may be any appropriate magnetic storage medium or magneto-optical storage medium, such as a resistive random access memory RRAM (Resistive Random Access Memory), a dynamic random access memory DRAM (Dynamic Random Access Memory), a static random access memory SRAM (Static Random-Access Memory), an enhanced dynamic random access memory EDRAM (Enhanced Dynamic Random Access Memory), a high-bandwidth memory HBM (High-Bandwidth Memory), a hybrid memory cube HMC (Hybrid Memory Cube), etc., or any other medium that can be used to store the required information and can be accessed by an application, a module, or both. Any such computer storage medium may be part of a device or accessible or connectable to a device. Any application or module described in the present invention may be implemented using computer-readable / executable instructions that may be stored or otherwise maintained by such a computer-readable medium.

[0103] In the description of this specification, "plurality" or "several" means at least two, such as two, three or more, etc., unless otherwise clearly and specifically defined.

[0104] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.

Claims

1. A laser treatment device control method based on optical imaging guidance, characterized in that: include: Collect fundus images, use a seed growth algorithm to determine the positions of blood vessel pixels in the fundus images, and then use the positions of blood vessel pixels to adjust the parameters of the laser treatment device, wherein using the seed growth algorithm to determine the positions of blood vessel pixels in the fundus images includes: Connected area detection is performed on the fundus image to obtain an approximate vascular connected domain; the selection value of each approximate vascular connected domain is calculated, and the pixel point with the smallest grayscale value in the approximate vascular connected domain whose selection value is greater than a preset selection value threshold is used as a seed point, and the pixel points in the neighborhood of the seed point are used as points to be grown; the selection value is negatively correlated with the grayscale mean of the edge pixel points of the corresponding connected domain and the gradient mean of its internal pixel points; the calculation expression of the selection value is: ; In the formula, represents the selected value of the i-th approximate vascular connectivity domain, represents the maximum distance between pixels in the i-th approximate vascular connected domain, represents the grayscale mean of the edge pixels of the i-th approximate vascular connected domain, Represents the gradient mean of the internal pixels of the i-th approximate vascular connectivity domain; The possibility that the point to be grown is a blood vessel pixel is calculated based on the deviation between the approximation between the neighboring pixel points of the point to be grown and the approximation between the neighboring pixel points of the pixel points in its growth path, and the possibility is negatively correlated with the deviation; the approximation between the neighboring pixel points of a certain pixel point is positively correlated with the gradient direction consistency of the pixel point and its related pixel points, and the related pixel points are located in the gradient direction of the pixel point and the distance from the pixel point is less than a preset threshold; the calculation expression of the possibility that the point to be grown is a blood vessel pixel point is: ; In the formula, Indicates the possibility that the jth growth point is a blood vessel pixel. represents the cosine of the angle between the line connecting the jth growth point to be grown and the previous growth point, and the line connecting the jth growth point to be grown and the pth pixel point in its growth path; represents the approximation between the neighboring pixels of the jth point to be grown, Indicates the approximation between the neighboring pixels of the pth pixel in the growth path of the jth point to be grown, is the maximum and minimum normalization function; Calculating the approximation between neighboring pixel points of a certain point to be grown includes: taking pixel points in the gradient direction of the point to be grown in a window centered on the point to be grown as relevant pixel points of the point to be grown, recording the number of pixel points in the window that are located in the gradient direction of the point to be grown and whose grayscale difference with the point to be grown is less than a preset grayscale difference value as A; recording the consistency of the gradient direction of the point to be grown and its related pixel points as B, calculating the approximation between the point to be grown and its related pixel points based on B and A, and recording it as E, and then calculating the approximation between the neighboring pixel points of the point to be grown based on E, the approximation between the neighboring pixel points of the point to be grown is positively correlated with E, and E is positively correlated with both B and A; The points to be grown whose likelihood is greater than a preset likelihood threshold are grown.

2. The method for controlling laser therapy equipment based on optical imaging guidance according to claim 1, characterized in that: Performing connected area detection on the fundus image includes: Gray-scaling the fundus image to obtain a fundus grayscale image; The collected fundus grayscale image is threshold segmented with a preset grayscale threshold to obtain a binary image, and the connected domain in the binary image is used as the approximate vascular connected domain.

3. The method for controlling laser therapy equipment based on optical imaging guidance according to claim 1, characterized in that: The possibility threshold is 0.5, and the selection value threshold is 0.

8.

4. The method for controlling laser therapy equipment based on optical imaging guidance according to claim 1, characterized in that: The grayscale difference preset value is 5.

5. The method for controlling laser therapy equipment based on optical imaging guidance according to claim 1, characterized in that: The calculation expression of the approximation between the point to be grown and its related pixel point is: ; In the formula, Indicates the approximation between the jth point to be grown and its related pixels, Indicates the number of related pixels of the jth growth point j, The unit vector representing the gradient direction of the nth related pixel of the jth point to be grown, Represents the unit vector of the gradient direction of the jth growth point, Represents the number of first approximate grayscale pixels, Represents the grayscale difference between the jth point to be grown and its nth related pixel.

6. The laser therapy device control method based on optical imaging guidance according to any one of claims 1 to 5, characterized in that: The calculating of the approximation between the neighboring pixels of a certain point to be grown also includes: The pixel points in the gradient direction of the previous growth point of the growth point to be grown in the window are recorded as the relevant pixel points of the previous growth point; the number of pixel points whose grayscale difference with the previous growth point is less than the preset grayscale difference value is recorded as C; The gradient direction consistency between the previous growth point and its related pixel points is recorded as D. The approximation between the previous growth point and its related pixel points is calculated based on C and D and recorded as F. The product of E and F is used as the approximation between the neighborhood pixel points of the growth point to be grown; F is positively correlated with both C and D.

7. A laser treatment system, characterized in that: It comprises an optical imaging device, a laser treatment device and a control system, wherein the optical imaging device is used to acquire fundus images, the laser treatment device is used to apply laser treatment to the eye, and the control system is used to execute the laser treatment device control method based on optical imaging guidance as described in any one of claims 1 to 6.