Vessel tortuosity evaluation method and system based on local-global ratio

Abstract

The embodiment of the present application provides a kind of based on local overall ratio's blood vessel tortuosity evaluation method and system, belong to image processing technical field.The method includes: the image information containing blood vessel is collected, and based on blood vessel path fitting result the blood vessel is split into multiple local blood vessels;The local blood vessel curvature of each local blood vessel is calculated respectively, and the tortuosity of current blood vessel is evaluated based on blood vessel path fitting result;Based on the local blood vessel curvature of all local blood vessels and the tortuosity of current blood vessel, the curvature of current blood vessel is obtained, and the tortuosity of current blood vessel is calculated based on the tortuosity of current blood vessel;Based on the curvature of current blood vessel and the tortuosity of current blood vessel, current tortuosity evaluation is carried out.The present application scheme solves the problems of existing blood vessel tortuosity evaluation method, such as qualitative one-sidedness, not enough objective, dependent on single index and the like.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of image processing, in particular to a blood vessel tortuosity evaluation method based on local-global ratio and a blood vessel tortuosity evaluation system based on local-global ratio. BACKGROUND

[0002] The retinal blood vessels will cause regional changes in the shape of the blood vessels and abnormal tortuosity of the blood vessels during the disease period of some patients, so how to effectively and accurately define, measure and analyze the tortuosity of the retinal fundus blood vessels is a long-term research difficulty. The accuracy of the evaluation result of the blood vessel tortuosity is essential for truly realizing clinical application. Traditionally, some methods based on the curvature integral of the blood vessel centerline are used to calculate the local blood vessel tortuosity, and the main disadvantage of this method is that two blood vessels with the same chord length and curve length but different curve numbers cannot be well distinguished. In view of the problem of low accuracy of the existing blood vessel tortuosity evaluation method, a new blood vessel tortuosity evaluation method needs to be created. SUMMARY

[0003] The purpose of the embodiments of the present application is to provide a blood vessel tortuosity evaluation method based on local-global ratio, so as to at least solve the problem of low accuracy of the existing blood vessel tortuosity evaluation method.

[0004] In order to achieve the above-mentioned purpose, the first aspect of the present application provides a blood vessel tortuosity evaluation method based on local-global ratio, which comprises: collecting image information containing blood vessels, and splitting the blood vessels into a plurality of local blood vessels based on blood vessel path fitting results; calculating the local blood vessel curvature of each local blood vessel respectively, and evaluating the tortuosity of the current blood vessel based on the blood vessel path fitting results; obtaining the curvature of the current blood vessel based on the local blood vessel curvature of all local blood vessels and the tortuosity of the current blood vessel, and calculating the tortuosity of the current blood vessel based on the tortuosity of the current blood vessel; and performing current tortuosity evaluation based on the curvature of the current blood vessel and the tortuosity of the current blood vessel.

[0005] Optionally, the splitting of the blood vessels into a plurality of local blood vessels based on the blood vessel path fitting results comprises: performing path fitting on the blood vessels in the image to obtain a complete blood vessel path; and dividing the complete blood vessel path into a plurality of path sub-segments, each path sub-segment being a local blood vessel.

[0006] Optionally, the calculating the local vessel curvature of each local vessel respectively comprises: presetting a plurality of discrete points along the local vessel, and drawing a circle with each discrete point as the center and with 1 / 2 of the length of the line connecting the two end points of the corresponding local vessel as the radius; each circle drawn corresponding to each discrete point divides the corresponding local vessel into two parts, and in the same local vessel, the part on one side of the corresponding local vessel is defined as region 1 and the part on the other side is defined as region 2; in the same local vessel, the area of region 1 or the area of region 2 of the circle drawn corresponding to each discrete point is identified as the split area of the circle drawn corresponding to each discrete point; and the curvature at the position of each discrete point is calculated based on the split area of the circle drawn corresponding to each discrete point, and the calculation formula is:

[0007]

[0008] wherein, k i is the curvature at the position of the i-th discrete point; b is the radius of the circle drawn corresponding to each discrete point of the current local vessel; S 1i is the split area of the circle drawn corresponding to the i-th discrete point; S is the total area of the circle drawn corresponding to the i-th discrete point; and the average vessel curvature of the corresponding local vessel is calculated based on the curvature of each discrete point of each local vessel, as the local vessel curvature of the corresponding local vessel.

[0009] Optionally, after the local vessel curvature of each local vessel is calculated respectively, the method comprises: calculating the average vessel curvature of the current vessel, and the calculation formula is:

[0010]

[0011] wherein, T1 is the average vessel curvature of the current vessel; and n is the number of discrete points.

[0012] Optionally, the evaluating the tortuosity of the current vessel based on the fitting result of the vessel path comprises: counting the curvatures of all local vessels and counting the number of changes in the sign of the curvature; wherein each change in the sign of the curvature is regarded as one bending of the vessel at the current position; the current vessel is divided into a plurality of circular arc segments based on the midpoint line between each adjacent two bending positions; wherein the vessel in each circular arc segment bends in the same direction only once; the arc length and the chord length of each circular arc segment are counted, and the tortuosity of the current vessel is evaluated based on the number of bendings of the current vessel and the arc length and the chord length of each circular arc segment.

[0013] Optionally, the obtaining the curvature of the current vessel based on the local vessel curvature of all local vessels and the tortuosity of the current vessel comprises: calculating the curvature density of the current vessel based on the tortuosity of the current vessel; wherein the calculation formula of the curvature density of the current vessel is:

[0014]

[0015] wherein, L d is the total length of the current vessel centerline; m is the number of bends of the current vessel; j is the jth circular arc segment; L xcj is the arc length of the jth circular arc segment; L xsj is the chord length of the jth circular arc segment; the curvature of the current vessel is calculated based on the average vessel curvature and the curvature density of the current vessel, and the calculation formula is:

[0016] M1 = T1 * T2

[0017] wherein, M1 is the curvature of the current vessel; T1 is the average local vessel curvature of the current vessel; and T2 is the curvature density of the current vessel.

[0018] Optionally, the tortuosity of the current vessel is calculated based on the bending condition of the current vessel, and the calculation formula of the tortuosity of the current vessel is:

[0019]

[0020] wherein, M2 is the tortuosity of the current vessel; m is the number of bends of the current vessel; k is the number of circular arc segments of the current vessel; θ is the curvature angle of each circular arc segment; and L is the arc length of each circular arc segment.

[0021] ca L xa is the chord length of each circular arc segment; and L t is the total length of the current vessel path fitting result.

[0022] Optionally, the current bending degree is evaluated based on the curvature of the current vessel and the tortuosity of the current vessel, including: an evaluation index of the current vessel is calculated based on the curvature of the current vessel and the tortuosity of the current vessel, and the calculation formula is:

[0023]

[0024] wherein, W is the evaluation index of the current vessel; M1 is the curvature of the current vessel; and M2 is the tortuosity of the current vessel; when the evaluation index of the current vessel is equal to 0, it indicates that the current vessel is a standard straight line; the greater the evaluation index of the current vessel, the greater the change range of the concave-convex state of the current vessel.

[0025] The second aspect of the present application provides a local-global ratio based vessel tortuosity evaluation system, the system comprising: an acquisition unit configured to acquire image information containing a blood vessel and split the blood vessel into a plurality of local blood vessels based on a blood vessel path fitting result; a processing unit configured to: calculate a local blood vessel curvature of each local blood vessel respectively, and evaluate a tortuosity of a current blood vessel based on the blood vessel path fitting result; obtain a curvature of the current blood vessel based on the local blood vessel curvatures of all the local blood vessels and the tortuosity of the current blood vessel, and calculate a tortuosity of the current blood vessel based on the tortuosity of the current blood vessel; and an evaluation unit configured to perform a current tortuosity evaluation based on the curvature of the current blood vessel and the tortuosity of the current blood vessel.

[0026] In another aspect, the present application provides a computer readable storage medium having instructions stored thereon, which, when executed on a computer, cause the computer to perform the above-mentioned local-global ratio based vessel tortuosity evaluation method.

[0027] Through the above technical solution, the basic idea of the present application is a local-global ratio based retinal tortuosity composite measurement definition and method, the present application introduces three indexes of average local blood vessel curvature, average local blood vessel curvature density and global blood vessel tortuosity consistency index to construct local tortuosity and global tortuosity, and the final ratio is defined as a new vessel tortuosity. It solves the problems of existing vessel tortuosity evaluation methods, such as qualitative one-sidedness, lack of objectivity, and dependence on a single index.

[0028] Other features and advantages of the present application will be described in detail in the following detailed description. BRIEF DESCRIPTION OF DRAWINGS

[0029] The accompanying drawings are included to provide a further understanding of the present application and constitute a part of the specification, which together with the detailed description, serve to explain the present application. In the drawings:

[0030] Figure 1 is a step flow chart of a local-global ratio based vessel tortuosity evaluation method provided by an embodiment of the present application;

[0031] Figure 2 is a schematic diagram of a principle of calculating each discrete point curvature provided by an embodiment of the present application;

[0032] Figure 3 is a schematic diagram of a blood vessel tortuosity provided by an embodiment of the present application;

[0033] Figure 4 is a schematic diagram of a local-global ratio based vessel tortuosity evaluation method implementation process provided by an embodiment of the present application;

[0034] Figure 5 is a system structure diagram of a blood vessel tortuosity evaluation system based on local-global ratio provided by an embodiment of the present application. DETAILED DESCRIPTION

[0035] The specific embodiments of the present application are described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are intended to illustrate and explain the present application, and are not intended to limit the present application.

[0036] The shape of the retinal blood vessels will change in the area due to the disease period of some patients, and the abnormal tortuosity of the blood vessels will occur, so how to effectively and accurately define, measure and analyze the shape of the retinal fundus blood vessels is a long-term research difficulty. The tortuosity of the retinal blood vessels plays an important role in the early diagnosis and treatment of eye diseases, and many diseases can be identified by checking the performance and image of the blood vessels, including diabetic retinopathy, retinopathy of prematurity, hypertensive retinopathy, etc. Badawi found that hypertensive retinopathy is caused by continuous high blood pressure leading to changes in the shape and diameter of the retinal blood vessels, and early detection of such changes can help prevent blindness and even death caused by stroke; Wang et al. also believe that the tortuosity of blood vessels as an indicator of retinal morphological changes can be quantitatively analyzed, and they proposed a blood vessel tortuosity analysis algorithm based on multiple subdivision, combined with a learning curve function of the number of blood vessel curvature inflection points, and emphasized human assessment.

[0037] In the existing method, considering the need to accurately define the related blood vessel tortuosity indicators among the shape, resolution, and size of the blood vessels, Brummer et al. rely on the numerical integration of the Frenet-Serret equation to enhance the blood vessel tortuosity measurement by reconstructing the three-dimensional blood vessel coordinates from the tortuosity measurement. Ramos L et al. also believe that the tortuosity of the retinal blood vessels can be used as a potential indicator of related blood and non-blood vessel diseases. They described a multi-specialist validation process for reference to calculate the tortuosity measurement of blood vessels, combined with a four-level scale of non-tortuous / tortuous, asymptomatic / symptomatic binary classification, and found that the prognostic performance of the tortuosity measurement was close to that of the experts.

[0038] However, the measured prognostic performance of the single classification or definition of the tortuosity of the blood vessels does not have mathematical accuracy and interpretability, and compared with the experience judgment relying on simple classification and expert assistance, the quantitative establishment of the corresponding tortuosity measurement index and the related measurement value can more comprehensively and accurately define the retinal blood vessel tortuosity. Joshi et al. developed a new quantitative measure of blood vessel tortuosity based on the curvature angle of the blood vessels, the length of the curved blood vessels on the chord length (the ratio of the arc to the chord), the number of changes in the curvature sign. Lisowska et al. also introduced distance metrics, tortuosity density, two curvature-based metrics, and the recently introduced slinky coding of general curves as five indicators of retinal blood vessel tortuosity, which standardize the selection of tortuosity index for clinical inference as much as possible. Turior et al. proposed a parallel algorithm based on the curvature calculated from the improved chain code algorithm, which uses a robust measure to quantify the retinal blood vessel tortuosity; Pourreza et al. also proposed a simple and efficient blood vessel tortuosity measurement method, which is roughly divided into blood vessel detection, extraction of blood vessel skeleton through thinning, detection of blood vessel intersections and bifurcations, and finally calculation of local and global tortuosity. However, these methods can only calculate the local or global tortuosity, and then analyze the two results respectively, which is not objective and lacks the overall definition of the blood vessel tortuosity, and the inference result is one-sided and dependent on a single indicator.

[0039] It can be seen that although there are some methods for evaluating the tortuosity of blood vessels at present, these methods still have the problem of low evaluation accuracy, and the accuracy of the evaluation result of the tortuosity of blood vessels is essential for truly realizing clinical application. In view of the problem of low accuracy of the existing blood vessel tortuosity evaluation method, the present application provides a blood vessel tortuosity evaluation method based on local-global ratio. The basic idea of the present application is a retinal tortuosity composite measurement definition and method based on local-global ratio. The present application introduces three indicators of average local blood vessel curvature, average local blood vessel curvature density and global blood vessel tortuosity consistency index to construct local tortuosity and global tortuosity, and the final ratio is defined as the new blood vessel tortuosity. It solves the problems of qualitative one-sidedness, lack of objectivity and dependence on a single indicator of the existing blood vessel tortuosity evaluation method.

[0040] Figure 1 is a method flowchart of the blood vessel tortuosity evaluation method based on local-global ratio provided by an embodiment of the present application. As shown in Figure 1 , the present application provides a blood vessel tortuosity evaluation method based on local-global ratio, which comprises the following steps:

[0041] Step S10: Collecting image information containing blood vessels, and splitting the blood vessels into a plurality of local blood vessels based on the blood vessel path fitting result.

[0042] Specifically, the scheme of the present application is to automatically identify the tortuosity of blood vessels through retinal images. In the existing method, the collected retinal images are processed to obtain the blood vessel network, which is relatively mature. The process of obtaining the blood vessel network on the retina through the retinal image is not described in detail. After identifying the blood vessels, the blood vessel path fitting is first performed based on the blood vessel network to find the path curve of the blood vessels. Because there are a large number of muscle blood vessels and arterial blood vessels on the retina, each blood vessel has a certain intersection, and if multiple blood vessels that do not belong to the same blood vessel are placed together for tortuosity evaluation, the intersection point may be incorrectly identified as the bending point of the corresponding blood vessel. Therefore, it is necessary to distinguish and path fit each blood vessel in order to accurately perform the subsequent steps.

[0043] Further, in order to facilitate image processing, the scheme of the present application further splits the blood vessel path fitting result after obtaining the image information containing the blood vessels. The obtained blood vessel path fitting result is split into a plurality of local blood vessels. Specifically, the blood vessels in the image are path fitted to obtain a complete blood vessel path. The complete blood vessel path is divided into a plurality of path subsegments, and each path subsegment is a local blood vessel. The scheme of the present application starts from the local blood vessel, which facilitates subsequent effective calculation of the average curvature and average curvature density of the local blood vessel, and overcomes the disadvantage that blood vessels with the same shape chord length and the same curve length but different curve numbers are not easy to distinguish.

[0044] In the embodiment of the present application, the traditional calculation of local blood vessel tortuosity adopts some methods based on the curvature integral of the blood vessel centerline. The main disadvantage of this method is that two blood vessels with the same chord length and curve length but different curve numbers cannot be well distinguished. In order to overcome the disadvantages of the traditional method, the scheme of the present application introduces three indexes of average local blood vessel curvature, average local blood vessel curvature density and overall blood vessel tortuosity consistency index to construct local tortuosity and overall tortuosity. The final ratio is defined as the new blood vessel tortuosity.

[0045] Step S20: Calculate the local blood vessel curvature of each local blood vessel respectively, and evaluate the tortuosity of the current blood vessel based on the blood vessel path fitting result.

[0046] Specifically, the calculation of the local vascular curvature of each local blood vessel includes: pre-setting multiple discrete points along the local blood vessel, and drawing a circle with each discrete point as the center and half the length of the line connecting the two ends of the corresponding local blood vessel as the radius; the circle corresponding to each discrete point is divided into two parts by the corresponding local blood vessel; within the same local blood vessel, the portion on one side of all corresponding local blood vessels is defined as region 1, and the portion on the other side is defined as region 2; within the same local blood vessel, the area of ​​region 1 or region 2 of the circles corresponding to all discrete points is simultaneously identified as the segmented area of ​​the circles corresponding to each discrete point; based on the identified segmented areas of the circles corresponding to each discrete point, the local curvature at the corresponding discrete point position is calculated, and the calculation formula is:

[0047]

[0048] Where, k i Let be the curvature of the i-th discrete point; b is the radius of the circle drawn corresponding to each discrete point of the current local blood vessel; S 1i S is the area of ​​the circle drawn corresponding to the i-th discrete point; S is the total area of ​​the circle drawn corresponding to the i-th discrete point; based on the curvature of each discrete point of each local blood vessel, the average blood vessel curvature of each local blood vessel is calculated and used as the local blood vessel curvature of each local blood vessel.

[0049] In embodiments of the present invention, such as Figure 2 To calculate the curvature of the curve at each point, a circle centered at a specified point and with radius b was drawn, where S is the total area of ​​the circle, and the value of b is chosen to be half the length of the line connecting the two endpoints of the local blood vessel. Figure 2 It can be seen that, under a certain ratio, the larger the ratio of S1 to S, the greater the curvature of the local blood vessel. When the chord length and arc length are equal, that is, when the area of ​​S1 is exactly 1 / 2 of S, the blood vessel basically mimics a straight line with a curvature of 0. Finally, the blood vessel curvature of all local blood vessel segments is the discrete average local blood vessel curvature T1, with a value range of [0, 3π / 2b).

[0050] Furthermore, after obtaining the local vascular curvature of each local blood vessel, it is necessary to evaluate the tortuosity of the current blood vessel based on the blood vessel path fitting results, count the curvature of all local blood vessels, and count the number of changes in the curvature sign; each change in the curvature sign indicates that the blood vessel has bent at the current position; based on the line connecting the midpoints between each two adjacent bending positions, the current blood vessel is divided into multiple arc segments; within each arc segment, the blood vessel bends in the same direction only once; the arc length and chord length of each arc segment are counted, and the tortuosity of the current blood vessel is evaluated based on the number of bends of the current blood vessel and the arc length and chord length of each arc segment.

[0051] In the embodiments of the present application, as Figure 3 Each time the bending occurs, the corresponding curvature sign changes once, so m is the number of curvature sign changes, that is, the change value of the second derivative of the center line, in common parlance, the number of concave-convex changes, that is, the number of inflection points of concave-convex.

[0052] In another possible implementation, the above-mentioned determination of the local blood vessel curvature is to arrange a plurality of discrete points along the blood vessel, and then make a circle on the discrete points, and determine the curvature of the blood vessel through the area ratio. Similarly, in the entire blood vessel, a circle with a fixed radius can also be preset, and the fixed radius is determined by the radii of the two end points of the minimum bending segment. Then the circle is moved along the blood vessel, and the area of the cutting region in the fixed direction is counted in real time. As it gets closer and closer to the concave point or the convex point, the area gets larger and larger or smaller and smaller, and as it gets farther and farther away from the concave point or the convex point, the area gets smaller and smaller or larger and larger. Based on the change rule of the area, it can be determined whether it passes through the concave-convex point, which can be realized as another possible implementation.

[0053] Step S30: obtaining the curvature of the current blood vessel based on the local blood vessel curvature of all local blood vessels and the bending condition of the current blood vessel, and calculating the tortuosity of the current blood vessel based on the bending condition of the current blood vessel.

[0054] Specifically, the curvature density is generally directly defined by the arc-chord ratio of the shape fitting. Generally, the distance measurement DM is used in the algorithm, as shown in the following formula:

[0055]

[0056] Where Lc is the length of the blood vessel center line, and Lx is the chord length (connecting the blood vessel end points). When the blood vessel is completely straight and as the tortuosity increases, DM is 1. However, DM cannot distinguish a blood vessel with multiple bends from a blood vessel with a single arc with the same average deviation chord, and the problem of this phenomenon is that DM cannot capture the local change of the global index. The curvature density itself is designed to discover the influence of the concave and convex tendencies of the local blood vessel on the local blood vessel curvature. In order to solve this phenomenon in order to better calculate the consistency of the concave-convex of the local blood vessel, the present application designs the following innovative curvature density formula to solve the problem:

[0057]

[0058] Where, L d is the total length of the current blood vessel center line; m is the number of bends of the current blood vessel; j is the jth arc segment; L xcj is the arc length of the jth arc segment; L xsjis the length of the chord of the jth arc segment. For a vessel with only one loop (i.e., a short vessel with only one concave-convex pattern), T2 = 0; for more than one loop, the tortuosity is greater than 0 (avoiding the problem of DM mentioned above), and finally the T2 index is also normalized to the vessel length (1 / L d ), which allows the comparison of vessels of various lengths and scales. The T2 index is more accurate than DM in comparing the tortuosity of vessels in simulated retinopathy images, with a value range of approximately [0, +∞).

[0059] After obtaining the curvature density, the curvature of the current vessel needs to be calculated based on the curvature density and the average vessel curvature of the vessel, and the calculation formula is:

[0060] M1 = T1 * T2

[0061] wherein M1 is the curvature of the current vessel; T1 is the average local vessel curvature of the current vessel; and T2 is the curvature density of the current vessel.

[0062] wherein the average tortuosity index formula of the curvature of n points is:

[0063]

[0064] wherein T1 is the average local vessel curvature of the current vessel; and n is the number of discrete points.

[0065] Further, the tortuosity of the current vessel is calculated based on the tortuosity of the current vessel, and the present scheme proposes a new overall vessel tortuosity consistency index, i.e., overall retinal vessel tortuosity, which is normalized and combined into a one-dimensional measure for representing the tortuosity of the vessel, and the calculation formula is:

[0066]

[0067] wherein M2 is the tortuosity of the current vessel; m is the number of bends of the current vessel; k is the number of arc segments of the current vessel; θ is the curvature angle size of each arc segment; and L is the arc length of each arc segment.

[0068] ca L xa is the length of the chord of the jth arc segment; and L t is the total length of the current vessel path fitting result. m is normalized on the total length of the vessel L t , so it is equal to the frequency per unit length but has the same tortuosity effect in the whole process of the vessel with different lengths. The vessel factor θ calculated for each segment is summed in all segments and normalized according to the number of each segment in the vessel. Because two vessels with approximately equal curvature but different lengths also have the same tortuosity performance. Finally, the arc-chord ratio is summed in all vessel segments and finally normalized according to the number of vessel segments divided in the whole.

[0069] Step S40: current tortuosity evaluation is performed based on the curvature of the current blood vessel and the tortuosity of the current blood vessel.

[0070] Specifically, the evaluation index of the current blood vessel is calculated based on the curvature of the current blood vessel and the tortuosity of the current blood vessel, and the calculation formula is as follows:

[0071]

[0072] Wherein, W is the evaluation index of the current blood vessel; M1 is the curvature of the current blood vessel; M2 is the tortuosity of the current blood vessel; and the evaluation index of the current blood vessel is equal to 0, indicating that the current blood vessel is a standard straight line; the greater the evaluation index of the current blood vessel, the greater the change range of the concave-convex state of the current blood vessel.

[0073]

[0074] In the embodiment of the present application, the calculation formula of the final blood vessel evaluation index is determined as follows:

[0075]

[0076]

[0077]

[0078] The evaluation index of the current blood vessel is explained as the ratio of the product of the discrete average blood vessel curvature and the average local curvature density to the overall retinal blood vessel tortuosity, and the value range is [0, +∞).

[0079] Further, the evaluation index of the current blood vessel provided in the present application can not only better distinguish the local blood vessels with the same frequency and equal arc-chord ratio, but also provide an objective and reproducible blood vessel tortuosity measurement method. In the finally obtained blood vessel tortuosity calculation formula, when the overall tortuosity is substantially consistent, the greater the average local blood vessel curvature density and the average local blood vessel curvature, that is, the more curved the concave-convex state of the local blood vessel and the local point, the greater the blood vessel tortuosity; when the local tortuosity, that is, the overall distribution of the average local blood vessel curvature and the average local blood vessel density, is substantially consistent, the more curved the overall tortuosity distribution, that is, the greater the overall blood vessel consistency index, the smaller the blood vessel tortuosity. When the result ratio is 0, it indicates the limit case that all blood vessels in the blood vessel network are standard straight lines; when the ratio is infinite, it indicates that the tortuosity of the blood vessel network is infinitely curved, that is, the concave-convex state of all blood vessels is extremely curved. In summary, the newly defined blood vessel tortuosity has good performance in general, and can better compare the newly defined blood vessel tortuosity from the aspects of local and overall tortuosity.

[0080] As Figure 4 ​The scheme can effectively calculate the local blood vessel average curvature and average curvature density, and overcome the disadvantage that blood vessels with the same shape chord length, the same curve length and different curve numbers are not easy to distinguish. A new blood vessel tortuosity calculation method is proposed by using the local tortuosity and overall tortuosity of the blood vessel, and the method can better play a role in disease prognosis and retinal related disease analysis and treatment. The time complexity is low, and the overall and local tortuosity of the blood vessel can be effectively and quickly identified and calculated.

[0081] Figure 5 The system structure diagram of the blood vessel tortuosity evaluation system based on a local overall ratio provided by an embodiment of the present application is shown in FIG. 1. Figure 5 As shown in FIG. 1, the embodiment of the present application provides a blood vessel tortuosity evaluation system based on a local overall ratio, which comprises: an acquisition unit configured to acquire image information containing blood vessels, and split the blood vessels into a plurality of local blood vessels based on a blood vessel path fitting result; a processing unit configured to: calculate the local blood vessel curvature of each local blood vessel respectively, and evaluate the tortuosity of the current blood vessel based on the blood vessel path fitting result; obtain the curvature of the current blood vessel based on the local blood vessel curvature of all local blood vessels and the tortuosity of the current blood vessel, and calculate the tortuosity of the current blood vessel based on the tortuosity of the current blood vessel; and an evaluation unit configured to perform current tortuosity evaluation based on the curvature of the current blood vessel and the tortuosity of the current blood vessel.

[0082] The embodiment of the present application further provides a computer readable storage medium, which stores instructions, and the instructions make the computer execute the blood vessel tortuosity evaluation method based on a local overall ratio when the computer runs.

[0083] Those skilled in the art can understand that all or part of the steps of the method in the above embodiments can be completed by programs instructing related hardware, the programs are stored in a storage medium, and the programs include a plurality of instructions for making a single-chip microcomputer, a chip or a processor execute all or part of the steps of the method described in each embodiment of the present application. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk and various program code storage media.

[0084] The optional embodiments of the present application are described in detail above with reference to the drawings, but the embodiments of the present application are not limited to the specific details in the above-described embodiments. Within the technical concept of the embodiments of the present application, various simple modifications can be made to the technical solutions of the embodiments of the present application, and these simple modifications all belong to the protection scope of the embodiments of the present application. In addition, it should be noted that, in the above-described specific embodiments, various specific technical features can be combined in any appropriate manner without contradiction. In order to avoid unnecessary repetition, various possible combinations are not described again in the embodiments of the present application.

[0085] In addition, various different embodiments of the present application can also be combined in any appropriate manner, as long as it does not deviate from the idea of the embodiments of the present application, and it should also be considered as disclosed by the embodiments of the present application.

Claims

1. A method for evaluating vessel tortuosity based on local-global ratio, characterized by, The method comprises: Collecting image information containing blood vessels, and splitting the blood vessels into a plurality of local blood vessels based on a blood vessel path fitting result; Respectively calculating the local blood vessel curvature of each local blood vessel, and evaluating the tortuosity of the current blood vessel based on the blood vessel path fitting result; wherein, The evaluation of the tortuosity of the current blood vessel based on the blood vessel path fitting result comprises: counting the curvatures of all local blood vessels, and counting the number of changes in the curvature sign; wherein, each change in the curvature sign is regarded as a bend of the blood vessel at the current position; the current blood vessel is divided into a plurality of circular arc segments based on the midpoint connecting line between each adjacent two bend positions; wherein, the blood vessel in each circular arc segment has and only has one bend in the same direction; the circular arc length and the circular arc chord length of each circular arc segment are counted, and the tortuosity of the current blood vessel is evaluated based on the number of bends of the current blood vessel and the circular arc length and the circular arc chord length of each circular arc segment; the curvature of the current blood vessel is obtained based on the local blood vessel curvatures of all local blood vessels and the tortuosity of the current blood vessel, comprising: calculating the curvature density of the current blood vessel based on the tortuosity of the current blood vessel; wherein, The curvature density calculation formula of the current blood vessel is: wherein, is the total length of the current vessel centerline; is the number of bends of the current vessel; is the jth circular arc segment; is the arc length of the jth circular arc segment; is the chord length of the jth circular arc segment; and the curvature of the current vessel is calculated based on the average vessel curvature and the curvature density of the current vessel, and the calculation formula is: wherein, is the curvature of the current vessel; is the average local vessel curvature of the current vessel; is the curvature density of the current vessel; The tortuosity calculation formula of the current blood vessel based on the tortuosity of the current blood vessel is: wherein, is the tortuosity of the current vessel; is the number of bends of the current vessel; is the number of circular arc segments of the current vessel; is the curvature angle of each circular arc segment; is the arc length of each segment of the circular arc segment; is the chord length of each segment of the circular arc segment; is the total length of the current vessel path fitting result; The curvature of the current blood vessel is obtained based on the local blood vessel curvatures of all local blood vessels and the tortuosity of the current blood vessel, and the tortuosity of the current blood vessel is calculated based on the tortuosity of the current blood vessel; The current tortuosity evaluation is performed based on the curvature of the current blood vessel and the tortuosity of the current blood vessel.

2. The method of claim 1, wherein, The splitting of the blood vessels into a plurality of local blood vessels based on the blood vessel path fitting result comprises: Performing path fitting on the blood vessels in the image to obtain a complete blood vessel path; Dividing the complete blood vessel path into a plurality of path sub-segments, each path sub-segment being a local blood vessel.

3. The method of claim 1, wherein, The calculation of the local blood vessel curvature of each local blood vessel comprises: A plurality of discrete points are preset along the local blood vessel, and a circle is drawn with each discrete point as the center and 1 / 2 of the length of the connecting line between the two ends of the corresponding local blood vessel as the radius; Each circle drawn by a discrete point is divided into two parts by the corresponding local blood vessel, and in the same local blood vessel, the part on one side of the corresponding local blood vessel is defined as region 1, and the part on the other side is defined as region 2; In the same local blood vessel, the areas of region 1 or region 2 of all circles drawn by the discrete points are identified as the segmentation areas of the circles drawn by the discrete points; The curvature at the position of each discrete point is calculated based on the identified segmentation areas of the circles drawn by the discrete points, and the calculation formula is: wherein, is the curvature of the ith discrete point location; radius of the circle drawn for the current local blood vessel for the discrete point; the partitioned area of the circle corresponding to the i-th discrete point pair drawn; the total area of the circles drawn for the i-th discrete point pair; The average blood vessel curvature of each local blood vessel is calculated based on the curvatures of the discrete points of each local blood vessel, and is taken as the local blood vessel curvature of the corresponding local blood vessel.

4. The method of claim 3, wherein, After the calculation of the local blood vessel curvature of each local blood vessel, the method comprises: calculating the average blood vessel curvature of the current blood vessel, and the calculation formula is: wherein, is the average vessel curvature of the current vessel; n is the number of discrete points.

5. The method of claim 1, wherein, The current tortuosity evaluation based on the curvature of the current blood vessel and the tortuosity of the current blood vessel comprises: The evaluation index of the current blood vessel is calculated based on the curvature of the current blood vessel and the tortuosity of the current blood vessel, and the calculation formula is: wherein, is an evaluation index of the current blood vessel; is the current curvature of the vessel; the tortuosity of the current blood vessel; When the evaluation index of the current blood vessel is equal to 0, it indicates that the current blood vessel is a standard straight line. The greater the evaluation index of the current blood vessel, the greater the change range of the concave-convex state of the current blood vessel.

6. A local-global ratio based vessel tortuosity assessment system, characterized by, The system is applied to the blood vessel tortuosity evaluation method based on local-global ratio as claimed in any one of claims 1-5, and the system comprises: a collection unit configured to collect image information containing a blood vessel, and split the blood vessel into a plurality of local blood vessels based on a blood vessel path fitting result; a processing unit configured to: calculate the local blood vessel curvature of each local blood vessel respectively, and evaluate the tortuosity of the current blood vessel based on the blood vessel path fitting result; obtain the curvature of the current blood vessel based on the local blood vessel curvature of all local blood vessels and the tortuosity of the current blood vessel, and calculate the tortuosity of the current blood vessel based on the tortuosity of the current blood vessel; an evaluation unit configured to perform current tortuosity evaluation based on the curvature of the current blood vessel and the tortuosity of the current blood vessel. 7.A computer readable storage medium, having instructions stored thereon, which when executed on a computer, cause the computer to perform the blood vessel tortuosity evaluation method based on local-global ratio as claimed in any one of claims 1-5.

Images