Method, device, computing equipment and storage medium for calculating blood flow reserve fraction based on intravascular imaging

Abstract

The present invention discloses a method, device, computing equipment and storage medium for calculating blood flow reserve fraction based on intravascular images, comprising: obtaining an image sequence for intravascular image segmentation; performing coarse and fine segmentation on the image sequence to determine the vascular cavity boundary; mapping the determined vascular cavity boundary to an axial image generated based on the image sequence; screening a suspected bifurcation point sequence by analyzing the boundary intensity value within the neighborhood range of each vascular cavity boundary point in the axial image; after mapping the suspected bifurcation point sequence to each frame of the original image, determining the fan-shaped area formed by the arc formed by the suspected bifurcation point sequence and the catheter center point, determining the catheter artifact according to the similarity of the fan-shaped area on the image sequence, and deleting the bifurcation point corresponding to the catheter artifact in the suspected bifurcation point to obtain a determined bifurcation point; after determining the bifurcation vessel according to the determined bifurcation point, constructing a vessel model in combination with the main vessel and the bifurcation vessel determined according to the vascular cavity boundary; and calculating the FFR according to the vessel model.

Description

Technical Field

[0001] The present invention belongs to the field of medical image processing, and in particular relates to a method, an apparatus, a computing device and a storage medium for calculating a blood flow reserve fraction based on intravascular images. Background Art

[0002] In current clinical practice, the diagnosis of coronary artery disease is mostly performed using invasive imaging methods, of which coronary angiography is the most popular and is currently considered the gold standard for cardiac catheterization and hemodynamic assessment. However, the photometric images produced by coronary angiography only depict the two-dimensional projection diameter of the lumen at a specific projection angle, without any geometric information about the lumen or plaque. As a result, recent coronary angiography is often accompanied by newer intravascular imaging technologies such as intravascular ultrasound (IVUS) and optical coherence tomography (OCT), which use ultrasound and near-infrared light to generate cross-sectional images of the coronary arteries, respectively. OCT images have extremely high image contrast and spatial resolution (10-20 microns), and their spatial resolution can often reach about 10 times that of IVUS images, so they can clearly characterize the morphology of atherosclerotic plaques and the status of stent implantation, including stent adhesion and stent coverage.

[0003] Parameters such as the minimum lumen diameter, minimum lumen area, and degree of stenosis obtained by analyzing OCT and IVUS images are of great significance for the diagnosis of coronary artery disease. Therefore, accurate segmentation of the coronary artery lumen is one of the main challenges faced in OCT and IVUS image analysis.

[0004] Although OCT images and IVUS images can effectively obtain information on vascular stenosis, IVUS and OCT do not have the function of measuring the fractional flow reserve (FFR) of coronary artery blood flow, and cannot judge whether there is functional stenosis in coronary artery blood flow. If the fractional flow reserve of coronary artery blood flow needs to be measured, it is necessary to use an invasive pressure guidewire and measure the coronary artery in a state of maximum hyperemia under the condition of intravenous administration of adenosine. This not only increases the complexity of the operation, but also increases the risk and cost for patients.

[0005] Patent application with publication number CN107977965A discloses a non-invasive calculation method for FFR of superior mesenteric artery dissection based on CT images, and patent application with publication number CN106650029A discloses a fast calculation method and system for bifurcation vessel pressure difference and FFR based on CFD simulation. Both are non-invasive methods for measuring coronary artery FFR, but neither is based on intravascular imaging.

[0006] Therefore, a non-invasive method for measuring FFR based on intravascular imaging (IVUS / OCT) is needed to reduce the risk and cost of diagnosis. Summary of the invention

[0007] Based on the above, the purpose of the present invention is to provide a method, apparatus, computing device and storage medium for calculating the blood flow reserve fraction based on intravascular images. After performing vessel segmentation based on the intravascular images, a vessel model is constructed based on the vessel segmentation results, and FFR is calculated based on the vessel model. In this way, FFR can be obtained without injecting vasodilators, thereby avoiding damage to patients caused by vasodilators.

[0008] In a first aspect, an embodiment provides a method for calculating blood flow reserve fraction based on intravascular images, comprising the following steps:

[0009] Acquiring an image sequence obtained by performing intravascular imaging analysis on a target coronary vessel;

[0010] Perform coarse and fine segmentation on the image series to determine the vascular lumen boundaries;

[0011] The image sequence is used to generate an axial image at fixed angles along the catheter, and the vascular lumen boundary determined by coarse and fine segmentation is mapped to the axial image to obtain multiple axial images marked with vascular lumen boundary points;

[0012] The suspected bifurcation point sequence is screened by analyzing the boundary intensity value in the neighborhood of each vascular lumen boundary point in the axial image marked with the vascular lumen boundary points;

[0013] After mapping the suspected bifurcation point sequence to each frame of the original image, the fan-shaped area formed by the arc formed by the suspected bifurcation point sequence and the catheter center point is determined, the catheter artifact is determined according to the similarity of the fan-shaped area on the image sequence, and the bifurcation point corresponding to the catheter artifact in the suspected bifurcation point is deleted to obtain the determined bifurcation point;

[0014] After determining the bifurcated blood vessels according to the determined bifurcation points, a blood vessel model is constructed by combining the main blood vessels and bifurcated blood vessels determined according to the vascular cavity boundary;

[0015] The fractional flow reserve was calculated based on the vascular model.

[0016] In one embodiment, the step of performing coarse and fine segmentation on the image sequence to determine the vascular lumen boundary comprises:

[0017] Rough segmentation process: After the image is converted into a polar coordinate format image, the boundary strength value is calculated according to the gray value of the pixel point. According to the boundary strength value of the pixel point, the dijksta minimum path algorithm is used to search for the vascular cavity boundary in the polar coordinate format image. The obtained vascular cavity boundary is smoothed and reflected to the image to obtain the rough segmentation result.

[0018] Fine segmentation process: With the lumen boundary point in the coarse segmentation result as the center, the original image is resampled along the external normal direction, and the Dijksta minimum path algorithm is used to search for the vascular lumen boundary in the resampled image. The obtained vascular lumen boundary is smoothed and reflected to the image to obtain the fine segmentation result;

[0019] After mapping the blood vessel cavity boundary in the fine segmentation result to the adjacent next frame original image as the coarse segmentation result, the fine segmentation process is performed, and the images in the image sequence are cycled in this way to achieve coarse and fine segmentation of the image sequence to determine the blood vessel cavity boundary.

[0020] In one embodiment, converting an image into a polar coordinate format image includes: constructing a polar coordinate system including an angle and a polar axis with the center of the image as the pole, taking a plurality of discrete pixel points on the polar axis according to a length interval, taking a plurality of discrete pixel points in the angle direction according to an angle interval, performing an interpolation operation on each discrete pixel point, and obtaining an image in a polar coordinate format.

[0021] In one embodiment, the boundary strength value is calculated according to the gray value of the pixel point using the following formula:

[0022] Grad(i,j)=a*(I(i+1,j)-I(i,j))+b*((I(i+2,j)-I(i-1,j))+(I(i+1,j-1)-I(i,j-1)) +(I(i+1,j+1)-I(i,j+1)))+c*((I(i+2,j-1)-I(i-1,j-1))+(I(i+2,j+1)-I(i-1,j+1)))

[0023] Among them, Grad(i,j) represents the boundary intensity value of the pixel at position (i,j), I(·) represents the grayscale value of the pixel, and a, b, and c are distance weight coefficients respectively; preferably, the value range of the distance weight coefficient a is 0.8-1.2; the value range of the distance weight coefficient b is 0.4-0.6; the value range of the distance weight coefficient c is 0.2-0.3.

[0024] In one embodiment, when the dijksta minimum path algorithm is applied to search for the lumen boundary in the polar coordinate format image, the inverse of the boundary intensity value is used as the consumption value of the pixel point;

[0025] In one embodiment, before applying the dijksta minimum path algorithm to search for the lumen boundary in the polar coordinate format image, the boundary strength values ​​that are less than the minimum limit value are normalized to the minimum limit value by setting a minimum value limit to achieve filtering preprocessing of the boundary strength values; then, based on the boundary strength values ​​after filtering preprocessing, the dijksta minimum path algorithm is applied to search for the lumen boundary in the polar coordinate format image.

[0026] In one embodiment, determining the catheter artifact based on the similarity of the fan-shaped regions on the image sequence includes:

[0027] Continuously track the fan-shaped areas on the image sequence. When the similarity of the fan-shaped areas in two adjacent frames of images is greater than the set similarity threshold, the fan-shaped areas are determined to be the same area. When the same area runs through the entire image sequence, it is considered to be a catheter artifact.

[0028] In one embodiment, the similarity of adjacent sector-shaped regions is calculated by integrating the area similarity, angle similarity and arc similarity of the sector-shaped regions.

[0029] In one embodiment, after the obtained axial image marked with the vascular lumen boundary points is binarized and the vascular lumen region is denoised, the boundary strength value within the neighborhood of each vascular lumen boundary point is analyzed to screen the suspected bifurcation point sequence;

[0030] In one embodiment, the calculating of the blood flow reserve fraction according to the vascular model includes: calculating the blood flow reserve fraction using a computational fluid dynamics (CFD) method based on the vascular model, or calculating the blood flow reserve fraction using a dimensionality reduction pressure drop model on the vascular model;

[0031] Among them, when the blood flow reserve fraction is calculated using the fluid dynamics method based on the vascular model, the flow distribution is based on the cross-sectional area of ​​the bifurcated blood vessel outlet, and the cross-sectional area is determined as follows:

[0032] Based on the segmentation boundaries of the main vessel and the bifurcation vessel on the bifurcation section, the midpoint of the opening is found among the bifurcation opening points on the boundary of the main vessel. Starting from the midpoint, the farthest point is found on the boundary of the bifurcation vessel, and the length L1 is recorded. In the same way, the lengths L2, L3, ..., Ln are obtained in the image sequence several frames before and after. Then, a straight line is fitted based on the axial distance between all the equal lengths and the image frames. The angle between the bifurcation vessel and the main vessel is calculated from the slope of the straight line. Finally, the cross-sectional area of ​​the bifurcation vessel is obtained based on the angle and the area of ​​the bifurcation vessel.

[0033] In a second aspect, an embodiment provides a device for calculating blood flow reserve fraction based on intravascular images, comprising:

[0034] An acquisition module, used to acquire an image sequence obtained by performing intravascular imaging analysis on a target coronary vessel;

[0035] A vascular lumen boundary determination module is used to perform coarse and fine segmentation on the image sequence to determine the vascular lumen boundary;

[0036] An axial image generation module is used to generate axial images along the catheter at fixed angles along the image sequence, and map the vascular cavity boundary determined by the coarse and fine segmentation to the axial image to obtain multiple axial images marked with vascular cavity boundary points;

[0037] A suspected bifurcation point screening module screens suspected bifurcation point sequences by analyzing the boundary strength value in the neighborhood of each vascular lumen boundary point in the axial image marked with the vascular lumen boundary points;

[0038] A bifurcation point confirmation module is used to map the suspected bifurcation point sequence to each frame of the original image, determine the fan-shaped area formed by the arc formed by the suspected bifurcation point sequence and the center point of the catheter, determine the catheter artifact according to the similarity of the fan-shaped area on the image sequence, and delete the bifurcation point corresponding to the catheter artifact in the suspected bifurcation point to obtain a confirmed bifurcation point;

[0039] A blood vessel model building module is used to determine the bifurcated blood vessel according to the determined bifurcation point, and then build a blood vessel model by combining the main blood vessel and the bifurcated blood vessel determined according to the vascular cavity boundary;

[0040] The blood flow reserve fraction calculation module is used to calculate the blood flow reserve fraction according to the blood vessel model.

[0041] In a third aspect, an embodiment provides a computing device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method for calculating the blood flow reserve fraction based on intravascular imaging as described in the first aspect is implemented.

[0042] In a fourth aspect, an embodiment provides a computer storage medium having a computer program stored thereon, characterized in that when the computer program is processed and executed, the method for calculating the blood flow reserve fraction based on intravascular imaging described in the first aspect is implemented.

[0043] The technical solution provided in the above embodiment has at least the following beneficial effects:

[0044] The image is segmented in two processes: coarse segmentation and fine segmentation. The fine segmentation result of the previous frame is used as the coarse segmentation result of the current frame, and the fine segmentation process is continued, which ensures the continuity of the overall segmentation result of the lumen boundary in the image sequence. In this way, for images with residual blood in the coronary lumen, fine segmentation is performed based on the fine segmentation result of the previous image without blood residue, which can greatly reduce the influence of residual blood on the lumen segmentation accuracy. Compared with the dynamic contour algorithm, it can significantly improve the segmentation accuracy of blood images.

[0045] The Dijkstra minimum path algorithm is used to search and identify the lumen boundary in the image. Compared with the region growing algorithm, it can accurately distinguish branched blood vessels and avoid inaccurate lumen boundary segmentation caused by overestimation of the lumen at the bifurcation opening of the blood vessel.

[0046] The boundaries of the vascular cavity are determined by performing coarse and fine segmentation on the image sequence; the suspected bifurcation point sequence is analyzed and screened by converting the image sequence into an axial image, and the catheter artifact is confirmed by judging the similarity of adjacent frames in the fan-shaped area as the suspected bifurcated blood vessel area, thereby determining the bifurcation point and the bifurcated blood vessel, and then constructing a vascular model through the main blood vessel and the bifurcated blood vessel determined according to the vascular cavity boundary. Compared with the traditional OCT image processing method, only segmenting the OCT target blood vessel for stenosis analysis ignores the influence of the bifurcated blood vessel on morphology, and using a single blood vessel for hemodynamic calculation ignores the shunting image of the bifurcated blood vessel on the blood flow, which does not conform to the law of conservation of flow. The method of the present invention can reconstruct the vascular model more accurately, and the method of finally obtaining the vascular FFR by hemodynamic calculation can meet the law of conservation of flow, and the calculated FFR is more accurate. At the same time, the FFR can be obtained without injecting vasodilators, avoiding damage to the patient caused by vasodilators. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0048] Figure 1 is a flow chart of a method for calculating blood flow reserve fraction based on intravascular images provided by an embodiment;

[0049] Figure 2 A flowchart of performing coarse and fine segmentation of an image sequence to determine the boundary of a blood vessel lumen is provided by an embodiment;

[0050] Figure 3 is a first frame of OCT image provided by an embodiment;

[0051] Figure 4 An embodiment provides an image of polar coordinate format converted from an OCT image;

[0052] Figure 5 An embodiment provides a method for searching and identifying lumen boundaries by using a Dijkstra minimum path algorithm in a polar coordinate format image;

[0053] Figure 6 is a smoothing result map of the lumen boundary provided by an embodiment;

[0054] Figure 7 A schematic diagram of mapping a smoothed lumen boundary to a first frame OCT image is provided in an embodiment;

[0055] Figure 8 is an original tenth OCT image provided by an embodiment;

[0056] Fig. 9 is a resampled image corresponding to the tenth frame of OCT image provided by an embodiment;

[0057] Fig.10 An embodiment provides a method for searching and identifying lumen boundaries in a resampled image using a Dijkstra minimum path algorithm;

[0058] Fig.11 A schematic diagram of mapping a lumen boundary to a tenth frame of OCT image provided by an embodiment;

[0059] Fig.12 is a schematic diagram of lumen boundary comparison between the ninth image and the tenth image provided by an embodiment;

[0060] Fig.13 is a schematic diagram of a segmentation result of a bifurcation OCT image provided by an embodiment;

[0061] Fig.14 is a schematic diagram of a segmentation result of an OCT image with blood residue provided by an embodiment;

[0062] Fig.15 is an axial image after binarization provided by an embodiment;

[0063] Fig.16 is an axial image after denoising provided by an embodiment;

[0064] Fig.17 is an axial view with suspected bifurcation points marked, provided by an embodiment;

[0065] Fig.18 is an OCT image with sector-shaped areas marked provided by an embodiment;

[0066] Fig.19 is a schematic diagram of a bifurcated blood vessel and a main blood vessel provided by an embodiment;

[0067] Fig. 20 is a schematic diagram of calculating the cross-sectional area of ​​a bifurcated blood vessel provided by an embodiment;

[0068] Fig.21 It is a schematic diagram of the structure of a device for calculating FFR based on intravascular images provided by an embodiment. DETAILED DESCRIPTION

[0069] To make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific implementation methods described herein are only used to explain the present invention and do not limit the scope of protection of the present invention.

[0070] Figure 1 FIG. 1 is a flow chart of a method for calculating blood flow reserve fraction based on intravascular images provided by an embodiment. Figure 1 As shown, the method for calculating blood flow reserve fraction based on intravascular images provided in the embodiment includes the following steps:

[0071] S101, acquiring an image sequence obtained by performing intravascular image analysis on a target coronary vessel.

[0072] The intravascular image analysis includes performing intravascular ultrasound (IVUS) or optical coherence tomography (OCT) on the target coronary vessel, and the obtained image sequence includes an IVUS image sequence or an OCT image sequence. In the embodiment, the image sequence is essentially a continuous cross-sectional image of a section of the target coronary vessel, and therefore, the coronary image sequence is composed of N continuous cross-sectional images, which are respectively recorded as I1, I2, I3...I N For each frame of image, two processes, coarse segmentation and fine segmentation, are performed to obtain the lumen boundary contour in the image.

[0073] S102, performing coarse and fine segmentation on the image sequence to determine the boundary of the blood vessel cavity.

[0074] like Figure 2 As shown, in the embodiment, the image sequence is segmented coarsely and finely to determine the boundary of the vascular cavity, including:

[0075] Rough segmentation process: After the first frame image is converted into a polar coordinate format image, the boundary strength value is calculated according to the gray value of the pixel point. According to the boundary strength value of the pixel point, the dijksta minimum path algorithm is applied to search for the vascular cavity boundary in the polar coordinate format image. The obtained vascular cavity boundary is smoothed and reflected to the image to obtain the rough segmentation result.

[0076] Fine segmentation process: With the lumen boundary point in the coarse segmentation result as the center, the original image is resampled along the external normal direction, and the Dijksta minimum path algorithm is used to search for the vascular lumen boundary in the resampled image. The obtained vascular lumen boundary is smoothed and reflected to the image to obtain the fine segmentation result;

[0077] After mapping the blood vessel cavity boundary in the fine segmentation result to the adjacent next frame original image as the coarse segmentation result, the fine segmentation process is performed, and the images in the image sequence are cycled in this way to achieve coarse and fine segmentation of the image sequence to determine the blood vessel cavity boundary.

[0078] Figure 3 The first frame of OCT image is given as an example. Figure 3 The coordinate system of the first frame OCT image shown is a rectangular coordinate system, which can be consistent with the actual cross section of the blood vessel. Based on this, in a possible implementation, converting the image into a polar coordinate format image includes: constructing a polar coordinate system including an angle and a polar axis with the center of the image as the pole, taking a plurality of discrete pixel points on the polar axis according to the length interval, taking a plurality of discrete pixel points in the angle direction according to the angle interval, performing an interpolation operation on each discrete pixel point, and obtaining a polar coordinate format image.

[0079] Exemplarily, the origin of the rectangular coordinate system of the image, that is, the center of the image, can be taken as the pole, and the mutually perpendicular x-axis and y-axis can be taken as the polar axis (or 0° direction) and 90° direction of the polar coordinate system respectively. In the polar axis direction, n1 discrete points are taken according to the length interval spa, and in the angle direction, n2 discrete points are taken according to the angle interval rad. Preferably, the size of spa is 1-4 pixel units, the size of n1 is 50-400, the size of rad is 1°-9°, and the size of n2 is 40-360. Then, bilinear interpolation or other interpolation methods are used to interpolate each discrete point to obtain the new pixel value corresponding to the discrete point, so as to obtain the polar coordinate format image. Figure 3 The OCT image shown is converted according to spa = 1.5, n1 = 200, rad = 3.6°, n2 = 100 to obtain Figure 4 The polar coordinate format image is shown.

[0080] After obtaining an image in polar coordinate format with a size of n1*n2, for a pixel at any position (i, j), its grayscale value is I(i, j) and the boundary strength is Grad(i, j). In general processing methods such as Roberts operator and Sobel operator, the boundary strength is defined as the absolute value of the grayscale gradient. However, in the present invention, the coronary OCT image or IVUS exhibits the characteristics of low grayscale values ​​in the lumen area and high grayscale values ​​in the vascular wall area. For this image, it is not accurate to directly define the absolute value of the grayscale gradient as the boundary strength. After experimental exploration, in the embodiment, the boundary strength is defined as the grayscale gradient in the polar axis direction, that is, the boundary strength value is calculated according to the grayscale value of the pixel point using the following formula:

[0081] Grad(i,j)=a*(I(i+1,j)-I(i,j))+b*((I(i+2,j)-I(i-1,j))+(I(i+1,j-1)-I(i,j-1)) +(I(i+1,j+1)-I(i,j+1)))+c*((I(i+2,j-1)-I(i-1,j-1))+(I(i+2,j+1)-I(i-1,j+1)))

[0082] Wherein, Grad(i,j) represents the boundary strength value of the pixel at position (i,j), I(·) represents the gray value of the pixel, and a, b, and c are distance weight coefficients. Preferably, the value range of a is 0.8–1.2; the value range of b is 0.4–0.6; and the value range of c is 0.2–0.3.

[0083] It can be seen from the above boundary strength value calculation formula that the boundary strength of any pixel point is determined by its adjacent 3*4 points. The larger the boundary strength value, the greater the possibility that it is a lumen boundary, and the smaller the boundary strength value, the smaller the possibility that it is a lumen boundary.

[0084] In the embodiment, when the dijksta minimum path algorithm is applied to search for the lumen boundary in the polar coordinate format image according to the boundary strength value of the pixel point, the inverse of the boundary strength value is used as the consumption value cost of the pixel point, that is, cost=1 / Grad.

[0085] It should be noted that, according to the boundary intensity value calculation formula, for some pixel points, their boundary intensity values ​​may be negative, and negative boundary intensity values ​​or too small negative boundary intensity values ​​cannot be well applied to the dijksta minimum path algorithm for lumen boundary search. Based on this, in an embodiment, before applying the dijksta minimum path algorithm to search for lumen boundaries in polar coordinate format images, by setting a minimum limit σ for the boundary intensity value, the boundary intensity values ​​that are less than the minimum limit value are normalized to the minimum limit value, that is, if Grad<σ, then let Grad=σ, so as to achieve filtering preprocessing of the boundary intensity value; then, based on the boundary intensity value after filtering preprocessing, the dijksta minimum path algorithm is applied to search for the lumen boundary in the polar coordinate format image, and the obtained lumen boundary is as follows: Figure 5 In the embodiment, the minimum limit value σ of the boundary strength value may be set to be 5-10.

[0086] After obtaining the lumen boundary, the lumen boundary is smoothed to a certain extent. Figure 5 The result of smoothing the lumen boundary is shown in Figure 6 As shown, then Figure 6 The smoothed lumen boundary shown is back-mapped to the OCT image to obtain Figure 7 The rough segmentation result is shown.

[0087] For fine segmentation of the coarse segmentation results, when resampling the coarse segmentation results, the sampling interval is set in the range of 0.5-1.0 pixel units, and each lumen boundary point is taken as the center, and resampling is performed along the external normal direction according to the sampling interval to obtain a resampled image. Then, the dijksta minimum path algorithm is applied to the resampled image to identify a finer lumen boundary. The finer lumen boundary is mapped back to the original first frame image after smoothing filtering, that is, the final segmentation result of the first frame image I1 segmentation.

[0088] For the coarse and fine segmentation process of each frame in the image sequence, in the image sequence, the space interval between two adjacent images is small, so the lumen boundary corresponding to the two adjacent images will not produce too much mutation. After obtaining the final lumen boundary of the fine segmentation of image I1, the final lumen edge is mapped to image I2, which is used as the coarse segmentation result of image I2. The same method as the fine segmentation of image I1 is applied to obtain the fine segmentation lumen boundary of image I2. Iterate segmentation in this way until the last image I N After the segmentation is completed, the final segmentation result of the entire coronary artery image sequence can be obtained for subsequent analysis and processing of morphological parameters such as the minimum lumen diameter, the minimum lumen area, and the degree of stenosis.

[0089] For example Figure 8 The tenth OCT image I shown10 , the final lumen boundary of the ninth frame OCT image I9 is ​​mapped to the tenth frame OCT image as the tenth frame OCT image I 10 The rough segmentation result of the tenth frame OCT image I 10 The rough segmentation result is resampled according to the sampling interval Δs = 0.8 to obtain Fig. 9 The resampled image is shown in the figure, and then the Dijkstra minimum path algorithm is used to search the resampled image to obtain the following Fig.10 The lumen boundary shown is mapped to the tenth frame OCT image I 10 , and get Fig.11 The fine segmentation result shown in FIG. 1 shows a comparison between the lumen boundary in the fine segmentation result and the lumen boundary in the ninth frame OCT image I9. Fig.12 shown.

[0090] The segmentation results of bifurcated OCT images and OCT images with residual blood using the above-mentioned FFR calculation method based on intravascular images are as follows: Fig.13 and Fig.14 As shown, analysis Fig.13 and Fig.14 As shown, it can be seen that the lumen boundary is clearly and accurately presented in the segmentation result.

[0091] S103, generating axial images along the catheter at fixed angles along the image sequence, and mapping the vascular cavity boundary determined by the coarse and fine segmentation to the axial image, so as to obtain a plurality of axial images marked with vascular cavity boundary points.

[0092] In the axial image, the marked vascular lumen boundary points are distributed on the upper and lower boundaries of the vascular lumen area, wherein the vascular lumen area is formed between the upper and lower vascular lumen boundaries. The axial image marked with vascular lumen boundary points is used to determine the bifurcation point and thus determine the bifurcated blood vessel.

[0093] S104, screening a suspected bifurcation point sequence by analyzing the boundary strength value within the neighborhood of each vascular lumen boundary point in the axial image marked with the vascular lumen boundary points.

[0094] In the embodiment, in order to improve the accuracy of screening suspected bifurcation point sequences, after the obtained axial image marked with vascular cavity boundary points is binarized and the vascular cavity area is denoised, the boundary strength value in the neighborhood of each vascular cavity boundary point is analyzed to screen the suspected bifurcation point sequence.

[0095] Taking OCT image as an example, the axial image of OCT image is binarized to obtain Fig.15 The binary image is shown in the figure, and then the segmented vascular cavity area is uniformly blackened to achieve the purpose of removing noise, and the following is obtained: Fig.16By analyzing the boundary strength value of the boundary point neighborhood range, it is determined whether the boundary point is close to the blood vessel wall and the suspected bifurcation position is screened out. Fig.17 The midpoint is a suspected bifurcation point.

[0096] S105, after mapping the suspected bifurcation point sequence to each frame of the original image, determine the fan-shaped area formed by the arc formed by the suspected bifurcation point sequence and the center point of the catheter, determine the catheter artifact according to the similarity of the fan-shaped area on the image sequence, and delete the bifurcation point corresponding to the catheter artifact in the suspected bifurcation point to obtain a determined bifurcation point.

[0097] Considering that the catheter artifacts appear as fan-shaped areas in the original image, and the catheter artifacts are usually similar in structure and continuous in space in the image, based on this feature, the catheter artifacts are determined and deleted. In the embodiment, the identified suspected bifurcation points are mapped to the original image, and the bifurcation areas are segmented, so that the suspected bifurcation areas in each frame of the image are identified. Based on this, the fan-shaped area formed by the arc formed by the suspected bifurcation point sequence and the catheter center point is determined, such as Fig.18 As shown, then, the catheter artifact is determined based on the similarity of the fan-shaped areas on the image sequence, including: continuously tracking the fan-shaped areas on the image sequence, when the similarity of the fan-shaped areas in two adjacent frames of images is greater than a set similarity threshold, the fan-shaped areas are determined to be the same area, and when the same area runs through the entire image sequence, the same area is considered to be a catheter artifact. In the embodiment, the similarity threshold is set to 0.8, that is, when the similarity exceeds 0.8, the fan-shaped areas are considered to be the same area.

[0098] In the embodiment, the similarity of adjacent sector-shaped regions is calculated by integrating the area similarity, angle similarity and arc similarity of the sector-shaped regions. The calculation formulas for the area similarity α, angle similarity β and arc similarity γ of the sector-shaped regions are respectively:

[0099]

[0100] Where s represents area, θ represents angle, and δ represents radian, then the similarity of adjacent sector areas is

[0101] After the catheter artifact is determined, the bifurcation point corresponding to the catheter artifact is deleted from the suspected bifurcation points to obtain a determined bifurcation point, which is used to determine the bifurcation vessel.

[0102] S106, after determining the bifurcated blood vessel according to the determined bifurcation point, a blood vessel model is constructed by combining the main blood vessel and the bifurcated blood vessel determined according to the vascular cavity boundary.

[0103] In the embodiment, a seed point is determined at the position of the bifurcation point, and region growing is performed starting from the seed point to segment the bifurcated blood vessel region. Then, the main blood vessel and the bifurcated blood vessel determined according to the vascular cavity boundary are merged to construct a blood vessel model, such as Fig.19 As shown, in the embodiment, the marchingcube algorithm can be used to construct the blood vessel model.

[0104] S107, after determining the bifurcated blood vessel according to the determined bifurcation point, calculate the cross-sectional area of ​​the bifurcated blood vessel outlet.

[0105] In the embodiment, according to obtaining the segmentation boundary of the main blood vessel and the segmentation boundary of the bifurcated blood vessel on the bifurcated cross section, the midpoint of the opening is found among the bifurcated opening points of the main blood vessel boundary, and the farthest position point is found on the bifurcated blood vessel boundary from the midpoint, and the length L1 is recorded. In the same way, the lengths L2, L3, ..., Ln are obtained in the previous and next frames of the image sequence, such as Fig. 20 As shown, a straight line can be fitted according to the axial distance between L1, L2, L3 and the image frames, and the angle between the bifurcated vessel and the main vessel is calculated from the slope of the straight line. Finally, the cross-sectional area of ​​the bifurcated vessel is obtained according to the angle θ and the area of ​​the bifurcated vessel.

[0106] S108, after distributing the flow rate using the cross-sectional area of ​​the bifurcated blood vessel outlet, the FFR is calculated according to the blood vessel model.

[0107] In an embodiment, the FFR may be calculated by using a fluid dynamics method based on a blood vessel model, or by using a dimension reduction pressure drop model on the blood vessel model.

[0108] When using the CFD method to calculate FFR, it is necessary to divide the coronary tree into three-dimensional grids, set flow parameters and boundary conditions, and use finite volume method, finite element method and other methods for numerical solution calculation. Among them, the flow parameters include: assuming that blood is an incompressible Newtonian fluid, and its density and viscosity coefficient are constants. Boundary conditions include one or more of blood flow inlet boundary conditions, blood flow outlet boundary conditions, and blood vessel wall boundary conditions.

[0109] Among them, the blood flow inlet boundary conditions include one or more of the blood flow inlet pressure boundary conditions, the blood flow inlet flow boundary conditions or the blood flow inlet velocity boundary conditions. In the blood flow inlet pressure boundary conditions, the inlet pressure is determined by the patient's specific blood pressure, which can be the patient's blood pressure measured by the catheter port pressure sensor during coronary angiography or coronary OCT imaging. In the blood flow inlet flow boundary conditions, the inlet flow can be calculated from the lumen volume of the coronary vascular model and the time for blood flow to pass through the blood vessel obtained from the coronary angiography image. In the blood flow inlet velocity boundary conditions, the inlet velocity can be calculated from the centerline length of the coronary vascular model and the time for blood flow to pass through the blood vessel obtained from the coronary angiography image.

[0110] The blood flow outlet boundary conditions include the blood flow outlet flow boundary conditions, or one or more of the circuit outlet boundary conditions simulated by the lumped parameter model (LPM) method. In the blood flow outlet flow boundary conditions, the outlet flow is determined by the area size relationship of each outlet, and satisfies the flow conservation with the blood flow inlet flow. In the blood flow outlet LPM boundary conditions, an external circuit model is connected at the outlet for simulation. The current represents the blood flow, the voltage represents the blood pressure, and the resistance represents the flow resistance from the outlet to the distal end of the coronary artery. The resistance size of each outlet is determined by the cross-sectional area of ​​each outlet.

[0111] The blood vessel wall boundary conditions include no-slip rigid wall boundary conditions. The rigid wall boundary conditions assume that the blood vessel wall does not deform during the calculation, which can greatly simplify the calculation process.

[0112] In one embodiment, the process of calculating FFR using the CFD method includes:

[0113] Step 1: Mesh the coronary artery model.

[0114] Step 2: Set the flow parameters. The blood flow density and viscosity coefficient are both set as constants.

[0115] Step 3: Set boundary conditions.

[0116] In the embodiment, the blood flow inlet pressure Pa is the patient's blood pressure measured by the catheter port pressure sensor during coronary angiography or coronary OCT imaging. The blood flow inlet flow rate Q is calculated from the lumen volume of the coronary vascular model and the time for blood flow to pass through the blood vessel obtained from the coronary angiography image. In the coronary angiography image, the corresponding image frame number when the blood flows through the inlet and outlet of the model is found, and the time t for blood flow to pass through the vascular model is calculated based on the frame rate of the angiography image. The volume V of the corresponding vascular segment model is calculated in the three-dimensional vascular model. The blood flow inlet flow rate Q = V / t.

[0117] In the embodiment, the blood flow outlet boundary condition is set as the LPM resistance boundary condition. The total resistance of the coronary model R = Pa / Q. The flow distribution is obtained according to the cross-sectional area S of each outlet of the coronary model, and the resistance of each outlet is calculated. The resistances of each outlet are connected in parallel. The flow distribution formula is Q~S n , which means that the blood flow rate Q at the entrance is proportional to the nth power of the cross-sectional area S, and n can take values ​​of 1 to 3 / 2.

[0118] In the embodiment, the blood vessel wall boundary is a no-slip rigid wall boundary condition, that is, it is considered that the blood vessel does not deform during the calculation, and the blood flow velocity at the wall is zero.

[0119] Step 4: Based on the above boundary conditions, the finite volume method is used to perform discrete solution calculations on the flow equation until the calculation converges, and the flow state of each part of the coronary artery is obtained, including the blood flow velocity and blood flow pressure Pd.

[0120] Step 5, calculate the coronary FFR value. According to the above blood flow inlet pressure Pa and the calculated blood flow pressure Pd at each coronary artery, the FFR value at that location is obtained: FFR = Pd / Pa.

[0121] When using the dimensionality reduction pressure drop model to calculate FFR, it is necessary to divide the center line of the coronary model and the three-dimensional vascular model into one-dimensional grids along the center line normal, set the flow parameters and boundary conditions, and use the dimensionality reduction pressure drop to perform numerical solution calculations. Among them, the flow parameters include: assuming that the blood is an incompressible Newtonian fluid, and its density and viscosity coefficient are both constants. The boundary conditions include: a given blood flow inlet pressure, a given blood flow inlet velocity or blood flow inlet flow, and a given blood flow outlet flow or blood flow outlet velocity at each outlet of the bifurcated coronary vascular model.

[0122] Among them, the blood flow inlet pressure is determined by the patient's specific blood pressure, which can be measured by the patient's blood pressure measured by the catheter port pressure sensor during coronary angiography or coronary OCT imaging. The blood flow inlet flow rate can be calculated from the lumen volume of the coronary vascular model and the time it takes for blood to pass through the blood vessel obtained from the coronary angiography image. The blood flow inlet velocity can be calculated from the centerline length of the coronary vascular model and the time it takes for blood to pass through the blood vessel obtained from the coronary angiography image. The blood flow outlet flow rate is determined by the relationship between the cross-sectional areas of each outlet and satisfies the flow conservation law with the blood flow inlet flow rate. The blood flow outlet velocity is calculated from the centerline length of the bifurcation segment of the coronary vascular model at the outlet and the time it takes for blood to pass through the bifurcation segment obtained from the coronary angiography image.

[0123] In one embodiment, the process of calculating FFR using the dimensionality reduction pressure drop model includes:

[0124] Step 1: Perform one-dimensional mesh division on the center line of the coronary artery model and the three-dimensional blood vessel model along the center line normal.

[0125] Step 2: Set the flow parameters. The blood flow density and viscosity coefficient are both set as constants.

[0126] Step 3: Set boundary conditions.

[0127] In the embodiment, the blood flow inlet pressure Pa is the patient's blood pressure measured by a catheter port pressure sensor during coronary angiography or coronary OCT imaging.

[0128] The blood flow entrance velocity u0 is calculated from the centerline length of the coronary vessel model and the time it takes for blood to pass through the vessel obtained from the coronary angiography image. In the coronary angiography image, find the corresponding image frame number when the blood flows through the model entrance and exit, and calculate the time t for blood to pass through the vessel model based on the frame rate of the angiography image. Calculate the length L of the centerline of the corresponding vessel segment in the three-dimensional vessel model. The blood flow entrance velocity u0 = L / t.

[0129] Blood flow outlet velocity u i The time t of blood flow passing through the bifurcation segment of the vascular model is calculated based on the centerline length of the bifurcation segment at the exit of the coronary artery model and the time t of blood flow passing through the bifurcation segment obtained from the coronary angiography image. In the coronary angiography image, the corresponding image frame number when the blood flow passes through the bifurcation point and exit of the bifurcation segment is found, and the time t of blood flow passing through the bifurcation segment of the vascular model is calculated based on the frame rate of the angiography image. i . Calculate the length L of the centerline of the corresponding vascular bifurcation segment in the three-dimensional vascular model i . Blood flow inlet velocity u i =L i / t i .

[0130] Step 4: Based on the above boundary conditions and the reconstructed three-dimensional coronary artery model and centerline, the coronary artery pressure drop is calculated using the dimensionality reduction pressure drop model. The pressure drops of the vascular segment and the vascular bifurcation segment are calculated separately, and finally the pressure drop ΔP at each point of the entire coronary artery model relative to the entrance of the coronary artery model can be obtained.

[0131] Step 5, calculate the coronary FFR value. According to the blood flow inlet pressure Pa and the calculated pressure drop ΔP at each coronary artery relative to the inlet of the coronary artery model, the FFR value at that location is obtained: FFR = (Pa-ΔP) / Pa.

[0132] Fig.21 The schematic diagram of the structure of the device for calculating FFR based on intravascular images provided in the embodiment is as follows: Fig.21 As shown, the device 2000 for calculating FFR based on intravascular images provided by the embodiment includes:

[0133] An acquisition module 2001 is used to acquire an image sequence obtained by performing intravascular image analysis on a target coronary vessel;

[0134] The vascular lumen boundary determination module 2002 is used to perform coarse and fine segmentation on the image sequence to determine the vascular lumen boundary;

[0135] An axial image generation module 2003 is used to generate axial images along the catheter at fixed angles along the image sequence, and map the vascular cavity boundary determined by the coarse and fine segmentation to the axial image to obtain multiple axial images marked with vascular cavity boundary points;

[0136] The suspected bifurcation point screening module 2004 screens the suspected bifurcation point sequence by analyzing the boundary strength value within the neighborhood of each vascular cavity boundary point in the axial image marked with the vascular cavity boundary point;

[0137] The bifurcation point confirmation module 2005 is used to map the suspected bifurcation point sequence to each frame of the original image, determine the fan-shaped area formed by the arc formed by the suspected bifurcation point sequence and the catheter center point, determine the catheter artifact according to the similarity of the fan-shaped area on the image sequence, and delete the bifurcation point corresponding to the catheter artifact in the suspected bifurcation point to obtain the determined bifurcation point;

[0138] The blood vessel model building module 2006 is used to build a blood vessel model by combining the main blood vessel and the bifurcated blood vessel determined according to the blood vessel cavity boundary after determining the bifurcated blood vessel according to the determined bifurcation point;

[0139] The FFR calculation module 2007 is used to calculate the FFR according to the blood vessel model.

[0140] It should be noted that the FFR calculation device based on intravascular images provided in the above embodiment should be illustrated by the division of the above functional modules when performing FFR calculation. The above functions can be assigned to different functional modules as needed, that is, the internal structure of the terminal or server is divided into different functional modules to complete all or part of the functions described above. In addition, the FFR calculation device based on intravascular images provided in the above embodiment and the FFR calculation method based on intravascular images are of the same concept. The specific implementation process is detailed in the FFR calculation method based on intravascular images, which will not be repeated here.

[0141] The method and device for calculating the blood flow reserve fraction based on intravascular images provided in the embodiment adopt two processes of coarse segmentation and fine segmentation for the image, and use the fine segmentation result of the previous frame image as the coarse segmentation result of the current frame image, and continue the fine segmentation process, thereby ensuring the continuity of the overall segmentation result of the lumen boundary in the image sequence; thus, for images with residual blood in the coronary lumen, fine segmentation is performed based on the fine segmentation result of the previous image without residual blood, which can greatly reduce the influence of residual blood on the lumen segmentation accuracy, and can significantly improve the segmentation accuracy of blood images compared with the dynamic contour algorithm;

[0142] The method and device for calculating the blood flow reserve fraction based on intravascular images provided in the embodiment use the Dijkstra minimum path algorithm to search and identify the lumen boundary in the image. Compared with the region growing algorithm, it can accurately distinguish branch blood vessels and avoid inaccurate lumen boundary segmentation caused by overestimating the lumen at the bifurcation opening of the blood vessel.

[0143] The method and device for calculating the blood flow reserve fraction based on intravascular images provided in the embodiment determine the vascular cavity boundary by performing coarse and fine segmentation on the image sequence; analyze and screen the suspected bifurcation point sequence by converting the image sequence into an axial image, and confirm the catheter artifact by performing adjacent frame similarity judgment on the fan-shaped area as the suspected bifurcated blood vessel area, thereby determining the bifurcation point and the bifurcated blood vessel, and then constructing a vascular model by the main blood vessel and bifurcated blood vessel determined according to the vascular cavity boundary, so that the vascular model can be reconstructed more accurately. Finally, the method of obtaining the vascular FFR by using hemodynamic calculation can meet the law of conservation of flow, and the calculated FFR is more accurate. At the same time, the FFR can be obtained without injecting vasodilators, thereby avoiding damage to the patient caused by vasodilators.

[0144] The embodiment further provides a computing device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the above-mentioned method for calculating FFR based on intravascular images is implemented, that is, the following steps are implemented:

[0145] S101, acquiring an image sequence obtained by performing intravascular imaging analysis on a target coronary vessel;

[0146] S102, performing coarse and fine segmentation on the image sequence to determine the boundary of the blood vessel lumen;

[0147] S103, generating an axial image at fixed angles along the catheter along the image sequence, and mapping the vascular lumen boundary determined by the coarse and fine segmentation to the axial image, so as to obtain a plurality of axial images marked with vascular lumen boundary points;

[0148] S104, screening a suspected bifurcation point sequence by analyzing the boundary strength value in the neighborhood of each vascular lumen boundary point in the axial image marked with the vascular lumen boundary points;

[0149] S105, after mapping the suspected bifurcation point sequence to each frame of the original image, determine the fan-shaped area formed by the arc formed by the suspected bifurcation point sequence and the center point of the catheter, determine the catheter artifact according to the similarity of the fan-shaped area on the image sequence, and delete the bifurcation point corresponding to the catheter artifact in the suspected bifurcation point to obtain a determined bifurcation point;

[0150] S106, after determining the bifurcated blood vessel according to the determined bifurcation point, construct a blood vessel model by combining the main blood vessel and the bifurcated blood vessel determined according to the vascular cavity boundary;

[0151] S107, calculating FFR according to the vascular model.

[0152] In practical applications, the memory can be a proximal volatile memory such as RAM, or a non-volatile memory such as ROM, FLASH, floppy disk, mechanical hard disk, etc., or a remote storage cloud. The processor can be a central processing unit (CPU), a microprocessor (MPU), a digital signal processor (DSP), or a field programmable gate array (FPGA), that is, the automatic segmentation step of the coronary OCT image can be implemented by these processors.

[0153] The embodiment further provides a computer storage medium on which a computer program is stored. When the computer program is processed and executed, the above-mentioned method for calculating FFR based on intravascular images is implemented, that is, the following steps are implemented:

[0154] S101, acquiring an image sequence obtained by performing intravascular imaging analysis on a target coronary vessel;

[0155] S102, performing coarse and fine segmentation on the image sequence to determine the boundary of the blood vessel lumen;

[0156] S103, generating an axial image at fixed angles along the catheter along the image sequence, and mapping the vascular lumen boundary determined by the coarse and fine segmentation to the axial image, so as to obtain a plurality of axial images marked with vascular lumen boundary points;

[0157] S104, screening a suspected bifurcation point sequence by analyzing the boundary strength value in the neighborhood of each vascular lumen boundary point in the axial image marked with the vascular lumen boundary points;

[0158] S105, after mapping the suspected bifurcation point sequence to each frame of the original image, determine the fan-shaped area formed by the arc formed by the suspected bifurcation point sequence and the center point of the catheter, determine the catheter artifact according to the similarity of the fan-shaped area on the image sequence, and delete the bifurcation point corresponding to the catheter artifact in the suspected bifurcation point to obtain a determined bifurcation point;

[0159] S106, after determining the bifurcated blood vessel according to the determined bifurcation point, construct a blood vessel model by combining the main blood vessel and the bifurcated blood vessel determined according to the vascular cavity boundary;

[0160] S107, calculating FFR according to the vascular model.

[0161] In practical applications, the computer-readable storage medium may be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, and the like.

[0162] The specific implementation methods described above provide a detailed description of the technical solutions and beneficial effects of the present invention. It should be understood that the above is only the most preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, supplements and equivalent substitutions made within the scope of the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for calculating blood flow reserve fraction based on intravascular imaging, characterized in that: The following steps are involved: Acquiring an image sequence obtained by performing intravascular imaging analysis on a target coronary vessel; Performing coarse and fine segmentation on the image sequence to determine the vascular lumen boundary, including: a coarse segmentation process and a fine segmentation process, taking the fine segmentation result of the previous frame image as the coarse segmentation result of the current frame image, and continuing the fine segmentation process to determine the vascular lumen boundary; The image sequence is used to generate an axial image at fixed angles along the catheter, and the vascular lumen boundary determined by coarse and fine segmentation is mapped to the axial image to obtain multiple axial images marked with vascular lumen boundary points; The suspected bifurcation point sequence is screened by analyzing the boundary intensity value in the neighborhood of each vascular lumen boundary point in the axial image marked with the vascular lumen boundary points; After mapping the suspected bifurcation point sequence to each frame of the original image, the fan-shaped area formed by the arc formed by the suspected bifurcation point sequence and the catheter center point is determined, the catheter artifact is determined according to the similarity of the fan-shaped area on the image sequence, and the bifurcation point corresponding to the catheter artifact in the suspected bifurcation point is deleted to obtain the determined bifurcation point; After determining the bifurcated blood vessels according to the determined bifurcation points, a blood vessel model is constructed by combining the main blood vessels and bifurcated blood vessels determined according to the vascular cavity boundary; The blood flow reserve fraction is calculated according to the blood vessel model, including: calculating the blood flow reserve fraction based on the blood vessel model using a fluid mechanics method, distributing the flow based on the cross-sectional area of ​​the bifurcated blood vessel outlet, and determining the cross-sectional area in the following manner: finding the midpoint of the opening in the bifurcation opening point of the main blood vessel boundary based on the segmentation boundary of the bifurcated blood vessel and the segmentation boundary of the bifurcated blood vessel on the bifurcated cross section, finding the farthest point on the bifurcated blood vessel boundary from the midpoint, recording the length L1, and in the same manner, obtaining the lengths L2, L3, ..., Ln in several frames before and after the image sequence, and then fitting a straight line based on all the equal distances of the lengths and the axial distances between the image frames, calculating the angle between the bifurcated blood vessel and the main blood vessel based on the slope of the straight line, and finally obtaining the cross-sectional area of ​​the bifurcated blood vessel based on the angle and the area of ​​the bifurcated blood vessel.

2. The method for calculating blood flow reserve fraction based on intravascular imaging according to claim 1, characterized in that: The rough segmentation process is as follows: after converting the image into a polar coordinate format image, the boundary strength value is calculated according to the gray value of the pixel point, and the Dijksta minimum path algorithm is used to search for the vascular cavity boundary in the polar coordinate format image according to the boundary strength value of the pixel point. The obtained vascular cavity boundary is smoothed and reflected to the image to obtain the rough segmentation result; The fine segmentation process is as follows: taking the lumen boundary point in the rough segmentation result as the center, resampling the original image along the external normal direction, applying the Dijksta minimum path algorithm to search for the vascular lumen boundary in the resampled image, smoothing the obtained vascular lumen boundary and reflecting it to the image to obtain the fine segmentation result; After mapping the blood vessel cavity boundary in the fine segmentation result to the adjacent next frame original image as the coarse segmentation result, the fine segmentation process is performed, and the images in the image sequence are cycled in this way to achieve coarse and fine segmentation of the image sequence to determine the blood vessel cavity boundary.

3. The method for calculating blood flow reserve fraction based on intravascular imaging according to claim 2, characterized in that: The method of converting an image into a polar coordinate format image includes: constructing a polar coordinate system including an angle and a polar axis with the center of the image as the pole, taking a plurality of discrete pixel points on the polar axis according to a length interval, taking a plurality of discrete pixel points in the angle direction according to an angle interval, performing an interpolation operation on each discrete pixel point, and obtaining a polar coordinate format image.

4. The method for calculating blood flow reserve fraction based on intravascular imaging according to claim 2, characterized in that: The boundary intensity value is calculated based on the gray value of the pixel using the following formula: Grad(i,j)=a*(I(i+1,j)-I(i,j))+b*((I(i+2,j)-I(i-1,j))+(I(i+1,j-1)-I(i,j-1)) +(I(i+1,j+1)-I(i,j+1)))+c*((I(i+2,j-1)-I(i-1,j-1))+(I(i+2,j+1)-I(i-1,j+1))) Wherein, Grad(i,j) represents the boundary intensity value of the pixel at position (i,j), I(·) represents the gray value of the pixel, and a, b, and c are distance weight coefficients respectively; The distance weight coefficient a has a value range of 0.8-1.2; the distance weight coefficient b has a value range of 0.4-0.6; and the distance weight coefficient c has a value range of 0.2-0.

3.

5. The method for calculating blood flow reserve fraction based on intravascular imaging according to claim 2, characterized in that: When the dijksta minimum path algorithm is used to search for lumen boundaries in polar coordinate format images, the inverse of the boundary intensity value is used as the consumption value of the pixel point; Before applying the Dijksta minimum path algorithm to search for the lumen boundary in the polar coordinate format image, the boundary intensity values ​​that are less than the minimum limit value are normalized to the minimum limit value by setting a minimum value limit to achieve filtering preprocessing of the boundary intensity value; then, based on the boundary intensity value after filtering preprocessing, the Dijksta minimum path algorithm is applied to search for the lumen boundary in the polar coordinate format image.

6. The method for calculating blood flow reserve fraction based on intravascular imaging according to claim 1, characterized in that: The method of determining the catheter artifact according to the similarity of the fan-shaped areas on the image sequence includes: The fan-shaped areas on the image sequence are continuously tracked. When the similarity of the fan-shaped areas in two adjacent frames of images is greater than the set similarity threshold, the fan-shaped areas are determined to be the same area. When the same area runs through the entire image sequence, it is considered to be a catheter artifact.

7. The method for calculating blood flow reserve fraction based on intravascular imaging according to claim 6, characterized in that: The similarity of adjacent sector-shaped regions is calculated by combining the area similarity, angle similarity and arc similarity of the sector-shaped regions.

8. The method for calculating blood flow reserve fraction based on intravascular imaging according to claim 1, characterized in that: After binarization and denoising of the vascular lumen region on the axial images marked with vascular lumen boundary points, the boundary intensity values ​​in the neighborhood of each vascular lumen boundary point were analyzed to screen the suspected bifurcation point sequence; The calculating of the blood flow reserve fraction according to the vascular model includes: calculating the blood flow reserve fraction by using a dimensionality reduction pressure drop model on the vascular model.

9. A device for calculating blood flow reserve fraction based on intravascular imaging, characterized in that: include: An acquisition module, used to acquire an image sequence obtained by performing intravascular imaging analysis on a target coronary vessel; The vascular lumen boundary determination module is used to perform coarse and fine segmentation on the image sequence to determine the vascular lumen boundary, including: a coarse segmentation process and a fine segmentation process, taking the fine segmentation result of the previous frame image as the coarse segmentation result of the current frame image, and continuing the fine segmentation process to determine the vascular lumen boundary; An axial image generation module is used to generate axial images along the catheter at fixed angles along the image sequence, and map the vascular cavity boundary determined by the coarse and fine segmentation to the axial image to obtain multiple axial images marked with vascular cavity boundary points; A suspected bifurcation point screening module screens suspected bifurcation point sequences by analyzing the boundary strength value in the neighborhood of each vascular lumen boundary point in the axial image marked with the vascular lumen boundary points; A bifurcation point confirmation module is used to map the suspected bifurcation point sequence to each frame of the original image, determine the fan-shaped area formed by the arc formed by the suspected bifurcation point sequence and the center point of the catheter, determine the catheter artifact according to the similarity of the fan-shaped area on the image sequence, and delete the bifurcation point corresponding to the catheter artifact in the suspected bifurcation point to obtain a confirmed bifurcation point; A blood vessel model building module is used to determine the bifurcated blood vessel according to the determined bifurcation point, and then build a blood vessel model by combining the main blood vessel and the bifurcated blood vessel determined according to the vascular cavity boundary; The blood flow reserve fraction calculation module is used to calculate the blood flow reserve fraction according to the blood vessel model, including: calculating the blood flow reserve fraction based on the blood vessel model using a fluid mechanics method, and distributing the flow based on the cross-sectional area of ​​the bifurcated blood vessel outlet. The cross-sectional area is determined by: finding the midpoint of the opening in the bifurcation opening point of the main blood vessel boundary based on the segmentation boundary of the bifurcated blood vessel and the segmentation boundary of the bifurcated blood vessel on the bifurcated cross section, and starting from the midpoint, finding the farthest position point on the bifurcated blood vessel boundary, recording the length L1, and in the same way, obtaining the lengths L2, L3, ..., Ln in the image sequence before and after several frames, and then fitting a straight line based on all the equal distances of the lengths and the axial distances between the image frames, calculating the angle between the bifurcated blood vessel and the main blood vessel based on the slope of the straight line, and finally obtaining the cross-sectional area of ​​the bifurcated blood vessel based on the angle and the area of ​​the bifurcated blood vessel.

10. A computing device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method for calculating blood flow reserve fraction based on intravascular imaging as described in any one of claims 1 to 8 is implemented.

11. A computer storage medium having a computer program stored thereon, characterized in that: When the computer program is processed and executed, the method for calculating blood flow reserve fraction based on intravascular imaging according to any one of claims 1 to 8 is implemented.

Images