An ultrasonic speckle tracking blood flow velocity measurement method, system, device and medium

Abstract

The present invention discloses an ultrasonic speckle tracking blood flow velocity measurement method, system, device and medium, relating to the technical field of ultrasonic blood flow velocity measurement. The method includes: acquiring an ultrasonic image sequence of a vascular radial section, and determining a reference frame and a comparison frame; determining a plurality of tracking blocks in the reference frame, and determining a starting tracking block; determining a search area of the starting tracking block in the comparison frame, and determining the best matching block of the starting tracking block; taking the best matching block of the starting tracking block as a reference, on the one hand, sequentially determining the best matching blocks of all the tracking blocks above the starting tracking block in the order from bottom to top, and on the other hand, sequentially determining the best matching blocks of all the tracking blocks below the starting tracking block in the order from top to bottom; determining the blood flow velocity corresponding to each tracking block according to the pixel coordinates of all the tracking blocks and all the best matching blocks. The present invention can achieve accurate and rapid blood flow velocity measurement, and is not limited by the maximum detectable blood flow velocity.

Description

Technical Field

[0001] The present invention relates to the technical field of ultrasonic blood flow velocity measurement, and particularly to an ultrasonic speckle tracking blood flow velocity measurement method, system, device and medium. Background Art

[0002] With the rapid development of modern life and the changes in the national lifestyle, especially the acceleration of population aging and urbanization, the unhealthy lifestyle of residents has become increasingly prominent. The risk factors causing atherosclerosis have a more significant impact on the health of residents, such as tobacco use or second-hand smoke exposure, unreasonable dietary structure, insufficient physical activity, obesity, mental disorders, etc. The incidence of vascular diseases caused by atherosclerosis continues to increase. According to the "Report on Cardiovascular Health and Diseases in China 2020", so far, cardiovascular diseases have ranked first among the total causes of death of urban and rural residents. It is 46.66% in rural areas and 43.81% in urban areas. The economic burden brought by cardiovascular diseases to residents and society is increasing day by day.

[0003] Atherosclerosis is site-specific, and certain special arterial segments are more prone to atherosclerosis. As the main channel for cerebral blood supply, the bifurcation of the carotid artery is a high-incidence area of atherosclerosis. When blood flows through the external carotid artery, the wall shear stress decreases, resulting in a decrease in blood flow velocity, deformation of the velocity profile, and deposition of cholesterol and lipids in the blood, forming atherosclerosis. The occurrence of carotid atherosclerosis leads to insufficient cerebral blood supply and is prone to ischemic cerebrovascular diseases. The plaque shedding at this site is also extremely likely to cause cerebral infarction and stroke. Therefore, accurately assessing the risk of carotid atherosclerosis and the development of the disease course is crucial for reducing the high mortality and high incidence of cardiovascular diseases.

[0004] As described above, on the one hand, the change of the blood flow velocity profile is closely related to the development of the carotid atherosclerosis disease course. On the other hand, many hemodynamic parameters for quantitatively evaluating the health status of the vascular system, such as wall shear stress and velocity shear rate, can be calculated based on the blood flow velocity profile. Therefore, accurately detecting the blood flow velocity profile is helpful for the prevention and early detection of carotid atherosclerosis and is of great significance for the diagnosis and risk assessment of cardiovascular diseases.

[0005] Currently, in clinical measurement of blood flow velocity, ultrasonic examination is widely used to observe moving organs, such as the heart, blood vessels, etc., due to its advantages of low price, simple operation, non-invasive, non-radiative, rapid imaging, continuous dynamic and repeated scanning. The methods for measuring blood flow velocity using ultrasound are mainly divided into two categories:

[0006] One type is the ultrasonic diagnostic technique based on the Doppler effect. This method estimates the blood flow velocity by using the frequency shift or phase shift generated when red blood cells in the blood scatter or reflect sound waves. For example, continuous Doppler, pulsed Doppler, and color flow imaging. This type of velocity measurement method has the advantages of fast calculation speed and high accuracy, but is limited by the maximum detectable velocity and has an angular dependence.

[0007] Another type is the ultrasonic diagnostic technique for blood flow velocity measurement based on ultrasonic speckle tracking. This method estimates the blood flow velocity by using two adjacent B-mode ultrasound images. The process is as follows: First, define a reference frame and define a series of tracking blocks in the reference frame; Second, define the subsequent adjacent image as the comparison frame, define the search area of the tracking block in the comparison frame, define the matching block in the search area, and use the sum of squared errors algorithm formula to find the matching block with the minimum error from the tracking block as the best matching block; Second, calculate the displacement of the tracking block between the reference frame and the comparison frame; Finally, divide the displacement by the time interval between the two ultrasound images to obtain the blood flow velocity. Radially across the entire lumen, multiple tracking blocks can be tracked to achieve the measurement of the blood flow velocity field based on the ultrasound image. Here, four existing ultrasonic speckle tracking blood flow velocity measurement methods are taken as examples for a brief description.

[0008] First, the traditional ultrasonic speckle tracking blood flow velocity measurement method based on the global search method

[0009] Traditional ultrasonic speckle tracking blood flow velocity measurement method - global search method. The search strategy of the global search method is as Figure 1 , including the following steps: First, randomly select two adjacent B-mode ultrasound images from the B-mode ultrasound image sequence. The first frame image is used as the reference frame, and the second frame image is used as the comparison frame; In the reference frame, determine the position of the blood vessel lumen according to the echo amplitude, and define X tracking blocks from the upper tube wall to the lower tube wall; In the comparison frame, centered on the pixel coordinates of each tracking block, define X search areas for X tracking blocks, and the sizes of all search areas are the same; Next is the global search. The sum of squared errors algorithm is used to characterize the error, and the sum of squared errors of each block (i.e., Figure 1 the square marked as A) and the tracking block is calculated row by row and column by column in the entire search area. After obtaining all the sums of squared errors in the entire search area, calculate their minimum value, and use the matching block corresponding to the minimum value as the best matching block (i.e., Figure 1 the gray square marked as A), that is, search for the best matching block of the tracking block in a traversal manner; Finally, use the offset between the tracking block and its best matching block as the displacement, and divide it by the acquisition time difference between the reference frame and the comparison frame to calculate the blood flow velocity corresponding to the tracking block. This algorithm is simple and easy to understand, but has the following disadvantages: 1) Traversal search in the entire search area results in long calculation time; 2) Speckle decorrelation leads to low calculation accuracy; 3) Fixed-size search area results in a small measurement range of blood flow velocity, that is, the maximum detectable blood flow velocity is small.

[0010] Second, an improved ultrasound speckle tracking blood flow velocity measurement method based on a three-step search method

[0011] The difference between the improved ultrasound speckle tracking blood flow velocity measurement method based on the three-step search method and the traditional method lies in the different search strategies. Figure 2 . The following steps are included: First, randomly select two adjacent B-ultrasound image frames in the B-ultrasound image sequence, with the first frame as the reference frame and the second frame as the comparison frame; in the reference frame, determine the position of the vascular cavity according to the echo amplitude, and define X tracking blocks from the upper wall to the lower wall; in the comparison frame, define X search areas of X tracking blocks with the pixel coordinates of each tracking block as the center, and the sizes of all search areas are the same. The next step is a three-step search. In the first step, use a square as a template in the search area to define 9 points with a distance L between each pair (that is, Figure 2 9 blocks marked as A in the figure, 9 blocks form a square), take 9 blocks to be matched at 9 points, calculate the average sum of square errors between the 9 blocks to be matched and the tracking block, and find the minimum value of the 9 average sum of square errors (that is, Figure 2 The second step is to define 8 points L / 2 away from the point with the smallest average square error in the first step as the center and the square as the template (i.e., Figure 2 8 blocks marked as B in the figure, 8 blocks form a square), take 8 blocks to be matched at 8 points, calculate the average sum of square errors between the 8 blocks to be matched and the tracking block, and find the minimum value of the sum of square errors of the 8 average errors (that is, Figure 2 The third step is to define four points adjacent to the point with the smallest mean square error in the second step as the center (i.e., Figure 2 ), take 4 blocks to be matched at 4 points, calculate the average sum of square errors between the 4 blocks to be matched and the tracking block, and find the minimum value of the 4 average sums of square errors. The point with the minimum average sum of square errors found in the third step is used as the best matching block of the tracking block, and the offset between the tracking block and its best matching block is used as the displacement, which is divided by the acquisition time difference between the reference frame and the comparison frame to calculate the blood flow velocity corresponding to the tracking block. Compared with traditional ultrasound based on global search method, this algorithm can reduce the algorithm operation time and improve the calculation efficiency to a certain extent because it does not adopt a global traversal search strategy. However, when the step size is large during the search, it is easy to skip the best matching block, resulting in local optimal problems and inaccurate results.

[0012] Third, an improved ultrasound speckle tracking blood flow velocity measurement method based on diamond search method

[0013] The difference between the improved ultrasound speckle tracking blood flow velocity measurement method based on the diamond search method and the traditional method lies in the different search strategies.Figure 3 The diamond search (also known as the diamond search) algorithm has two different matching templates: a large diamond and a small diamond. The large diamond has 9 search points, while the small diamond has only 5 search points. First, a coarse search is performed using the large diamond search template with a larger step size, and then a fine search is performed using the small diamond template. It includes the following steps: First, randomly select two adjacent B-mode ultrasound images in the B-mode ultrasound image sequence. The first image is used as the reference frame, and the second image is used as the comparison frame. In the reference frame, determine the position of the blood vessel lumen based on the echo amplitude, and define X tracking blocks from the upper tube wall to the lower tube wall. In the comparison frame, centered on the pixel coordinates of each tracking block, define X search areas for the X tracking blocks, and all search areas have the same size. Next is the diamond search. In the first step, within the search area, use a diamond as the template to define 9 points of the diamond (i.e., Figure 3 the 9 squares marked as A in Figure 3 , and the 9 squares form a diamond. The distance from each vertex to the center point of the diamond is L). Take 9 blocks to be matched at the 9 points, calculate the sum of the mean squared errors between the 9 blocks to be matched and the tracking block, and find the minimum value of the 9 sums of the mean squared errors (i.e., Figure 3 the gray square marked as A in Figure 3 ); In the second step, if the center point of the search is the point with the minimum sum of the mean squared errors, then jump to the third step and use the small diamond search template. Otherwise, centered on the point with the minimum sum of the mean squared errors in the first step, use a diamond as the template to define 8 points with the same step size as in the first step. Among these 8 points, take the blocks to be matched at the remaining points except those for which the sum of the mean squared errors has already been calculated (i.e., Figure 3 the squares marked as B in Figure 3 ), calculate the sum of the mean squared errors between the blocks to be matched and the tracking block, and find the minimum value of the sum of the mean squared errors (i.e., Figure 3 the gray square marked as B in Figure 3 ), and repeat this step; In the third step, if the center point of the search is the point with the minimum sum of the mean squared errors (i.e., Figure 3 the gray square marked as B in Figure 3 ), then use the small diamond search template. The specific method is to define 4 points adjacent to the point with the minimum sum of the mean squared errors in the second step. Among these 4 points, take the blocks to be matched at the remaining points except those for which the sum of the mean squared errors has already been calculated (i.e., Figure 3 the squares marked as D in Figure 3the gray square marked as D in the figure). The point with the minimum sum of mean squared errors obtained in the third step is used as the best matching block of the tracking block, and the offset between the tracking block and its best matching block is used as the displacement, which is divided by the acquisition time difference between the reference frame and the comparison frame to calculate the blood flow velocity corresponding to the tracking block. Since this algorithm does not adopt a global traversal search strategy, compared with traditional ultrasound based on the global search method, it can reduce the algorithm operation time to a certain extent and improve the calculation efficiency. However, when the search step size is large, it is easy to skip the best matching block, resulting in the problem of local optimality and inaccurate results.

[0014] Fourth, the ultrasound speckle tracking blood flow velocity measurement method based on the optimal frame interval

[0015] The ultrasound speckle tracking blood flow velocity measurement method based on the optimal frame interval, as Figure 4 shown, includes the following steps: First, randomly select a frame as the reference frame 10 in the B-mode ultrasound image sequence, and select N frames as the comparison frame 12 sequence after the reference frame 10. Among them, the subsequent frame image of the reference frame 10 is the adjacent comparison frame 7; in the reference frame 10, determine the blood vessel lumen position according to the echo amplitude, and define X tracking blocks 11 in the blood vessel lumen; in the N comparison frames 12, with the pixel coordinates of each tracking block 11 as the center, define X×N search areas 9 of the X tracking blocks 11, and the sizes of all search areas 9 are the same; next, use the global search method to find the N best matching blocks of each tracking block 11 in the N search areas 9, and record the N cross-correlation coefficients; among the N cross-correlation coefficients, select the image that is less than the correlation threshold and the farthest from the reference frame 10 as the optimal frame interval. Use the optimal frame interval to calculate the blood flow velocity of the tracking block 11, and then determine the blood flow parabola 6, where the blood flow parabola 6 is determined by connecting the blood flow velocities of the blood flows at several different positions (such as the first blood flow 4, the second blood flow 5, and the maximum blood flow 6). This algorithm can accurately measure the slow blood flow velocity near the blood vessel wall 2, minimize the quantization error, and improve the measurement accuracy of the blood flow velocity. However, due to N global searches, the algorithm has a large amount of calculation and a long running time.

[0016] In summary, the current ultrasound speckle tracking blood flow velocity measurement still has the following three problems: 1) The traversal search leads to a long calculation time; 2) The speckle decorrelation leads to a low calculation accuracy; 3) The fixed-size search area leads to a small measurement range of the blood flow velocity, that is, the maximum detectable blood flow velocity is small. Summary of the Invention

[0017] The purpose of the present invention is to provide an ultrasound speckle tracking blood flow velocity measurement method, system, device and medium to achieve accurate and rapid blood flow velocity measurement and be not limited by the maximum detectable blood flow velocity.

[0018] To achieve the above object, the present invention provides the following solutions:

[0019] An ultrasonic speckle tracking blood flow velocity measurement method, comprising:

[0020] Obtaining an ultrasonic image sequence of a radial section of a blood vessel, and determining a reference frame and a comparison frame from the ultrasonic image sequence; the comparison frame is the subsequent frame image of the reference frame;

[0021] Determining a plurality of tracking blocks along the position where the radial blood flow is located in the reference frame, and determining any one of the tracking blocks in the middle area as the starting tracking block;

[0022] Determining a search area of the starting tracking block along the transverse direction inside the blood vessel in the comparison frame, and determining the best matching block of the starting tracking block within the search area of the starting tracking block;

[0023] Taking the best matching block of the starting tracking block as a reference, and sequentially determining the best matching blocks of all the tracking blocks above the starting tracking block in the comparison frame in the order from bottom to top;

[0024] Taking the best matching block of the starting tracking block as a reference, and sequentially determining the best matching blocks of all the tracking blocks below the starting tracking block in the comparison frame in the order from top to bottom;

[0025] Determining the blood flow velocity corresponding to each tracking block according to the pixel coordinates of all the tracking blocks and the pixel coordinates of all the best matching blocks respectively.

[0026] An ultrasonic speckle tracking blood flow velocity measurement system, comprising:

[0027] An image acquisition module, configured to obtain an ultrasonic image sequence of a radial section of a blood vessel, and determine a reference frame and a comparison frame from the ultrasonic image sequence; the comparison frame is the subsequent frame image of the reference frame;

[0028] A tracking block determination module, configured to determine a plurality of tracking blocks along the position where the radial blood flow is located in the reference frame, and determine any one of the tracking blocks in the middle area as the starting tracking block;

[0029] A best matching block determination module for the starting tracking block, configured to determine a search area of the starting tracking block along the transverse direction inside the blood vessel in the comparison frame, and determine the best matching block of the starting tracking block within the search area of the starting tracking block;

[0030] A best matching block determination module for the upper tracking blocks, configured to take the best matching block of the starting tracking block as a reference, and sequentially determine the best matching blocks of all the tracking blocks above the starting tracking block in the comparison frame in the order from bottom to top;

[0031] The optimal matching block determination module for the lower tracking block is configured to, with the optimal matching block of the starting tracking block as a reference, sequentially determine the optimal matching blocks of all the tracking blocks below the starting tracking block in the comparison frame in the order from top to bottom;

[0032] The blood flow velocity determination module is configured to determine the blood flow velocity corresponding to each tracking block according to the pixel coordinates of all the tracking blocks and the pixel coordinates of all the optimal matching blocks respectively.

[0033] An electronic device includes a memory and a processor. The memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to execute the above ultrasonic speckle tracking blood flow velocity measurement method.

[0034] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the above ultrasonic speckle tracking blood flow velocity measurement method is implemented.

[0035] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0036] Since blood has a certain viscosity and the blood flow velocity radial profile has a laminar flow characteristic, it results in two adjacent tracking blocks, and their optimal matching blocks are not far away. Therefore, in the ultrasonic speckle tracking blood flow velocity measurement method provided by the present invention, with the optimal matching block of the starting tracking block as a reference, the optimal matching blocks of all the tracking blocks above the starting tracking block are sequentially determined in the comparison frame in the order from bottom to top, and, in the order from top to bottom, the optimal matching blocks of all the tracking blocks below the starting tracking block are sequentially determined in the comparison frame. In this process, the position of the search area of each tracking block is determined according to the position of the optimal matching block of the adjacent tracking block, and there is no need to perform a large-area traversal search, but an extended search is performed around the position of the optimal matching block of the adjacent tracking block. Therefore, it can save the measurement time of the blood flow velocity, and consider the correlation of the blood flow velocity between each tracking block, avoid local outliers, and improve the measurement accuracy of the blood flow velocity; in addition, since there is no size limitation on the range of the search area of other tracking blocks except the starting tracking block, the search range can be infinitely expanded before the corresponding optimal matching block is found, so that the speckle tracking method is no longer limited by the maximum detectable velocity, and the measurement range of the blood flow velocity is expanded. Description of the Drawings

[0037] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention, and for those of ordinary skill in the art, other drawings can be obtained according to these drawings without creative efforts.

[0038] Figure 1 Schematic diagram of the traditional ultrasonic speckle tracking blood flow velocity measurement method based on the global search method;

[0039] Figure 2 Schematic diagram of the improved ultrasonic speckle tracking blood flow velocity measurement method based on the three-step search method;

[0040] Figure 3 Schematic diagram of the improved ultrasonic speckle tracking blood flow velocity measurement method based on the diamond search method;

[0041] Figure 4 Schematic diagram of the ultrasonic speckle tracking blood flow velocity measurement method based on the optimal frame interval;

[0042] Figure 5 Flow chart of the ultrasonic speckle tracking blood flow velocity measurement method provided by the present invention;

[0043] Figure 6 Specific flow chart of the ultrasonic speckle tracking blood flow velocity measurement method provided by the present invention;

[0044] Figure 7 Schematic diagram of the blood flow model of a specific embodiment of the present invention;

[0045] Figure 8 Schematic diagram of the reference frame provided by the present invention;

[0046] Figure 9 Schematic diagram of finding the best matching block of the starting tracking block provided by the present invention;

[0047] Figure 10 Schematic diagram of finding the best matching blocks corresponding to all tracking blocks provided by the present invention;

[0048] Figure 11 Schematic diagram of the mean square error sum value of the results obtained from a specific embodiment of the present invention;

[0049] Figure 12 Schematic diagram of the measured blood flow parabola result of a specific embodiment of the present invention.

[0050] Symbol description:

[0051] 1. Tissue; 2. Vessel wall; 3. Blood vessel; 4. First blood flow; 5. Second blood flow; 6. Maximum blood flow; 7. Adjacent comparison frame; 8. Blood flow parabola; 9. Search area; 10. Reference frame; 11. Tracking block; 12. Comparison frame. Detailed implementation manners

[0052] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0053] The object of the present invention is to provide an ultrasonic speckle tracking blood flow velocity measurement method, system, device and medium to achieve accurate and rapid blood flow velocity measurement and not be limited by the maximum detectable blood flow velocity.

[0054] To make the above objects, features and advantages of the present invention more obvious and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0055] Embodiment 1

[0056] The present invention provides an ultrasonic speckle tracking blood flow velocity measurement method. As Figure 5 and Figure 6 shown, it includes:

[0057] Step S1: Obtain an ultrasonic image sequence of the radial section of a blood vessel, and determine a reference frame and a comparison frame from the ultrasonic image sequence; the comparison frame is the subsequent frame image of the reference frame.

[0058] Step S2: Determine a plurality of tracking blocks along the position where the radial blood flow is located in the reference frame, and determine any one of the tracking blocks in the middle area as the starting tracking block.

[0059] Step S2 specifically includes: determining a plurality of tracking blocks along the position where the radial blood flow is located in the reference frame; numbering all the tracking blocks in the order from the upper blood vessel wall to the lower blood vessel wall; determining any one of the tracking blocks with the serial number between 0.4W and 0.6W as the starting tracking block; where W is the total number of tracking blocks.

[0060] Steps S1 and S2 are the algorithm initialization, which specifically includes the following steps:

[0061] First, use ultrasonic scanning to scan the radial section of the blood vessel to generate a time series of B-mode images. The radial section of the blood vessel is as Figure 7 shown, including tissue 1, blood vessel 3 and blood vessel wall 2, where V(d) represents the blood flow parabola formed due to the laminar flow characteristics of the blood flow velocity radial profile.

[0062] Secondly, randomly select a frame image in the time series of ultrasonic images of the radial section of the blood vessel as the reference frame, and use the subsequent frame image of the reference frame as the comparison frame.

[0063] After that, at the position of the radial blood flow in the reference frame, W windows with a pixel size of X×Y are selected as tracking blocks, as Figure 8 shown. Denote the pixel coordinates of the W tracking blocks in the reference frame as (H w , T w ), where 1 ≤ w ≤ W.

[0064] Finally, among the total W tracking blocks from the upper blood vessel wall to the lower blood vessel wall, select the tracking blocks with serial numbers between 0.4W and 0.6W as the starting tracking blocks. Let the serial number of the starting tracking block be i (the value of i is between 0.4W and 0.6W). The position of the starting tracking block is as Figure 8 shown by S(x, y) in

[0065] Step S3: Determine the search area of the starting tracking block in the transverse direction inside the blood vessel in the comparison frame, and determine the best matching block of the starting tracking block within the search area of the starting tracking block.

[0066] Step S3 specifically includes:

[0067] Step S3.1: Determine the search area of the starting tracking block in the transverse direction inside the blood vessel in the comparison frame.

[0068] Step S3.2: Determine several matching blocks of the starting tracking block within the search area of the starting tracking block; the matching blocks of the starting tracking block have the same size as the starting tracking block.

[0069] Step S3.3: Calculate the sum of the mean squared errors of all the matching blocks of the starting tracking block and the starting tracking block.

[0070] Step S3.4: Determine the matching block of the starting tracking block with the minimum sum of the mean squared errors from the starting tracking block as the best matching block of the starting tracking block.

[0071] Specifically, within the search area of the starting tracking block, find the best matching block of the starting tracking block as the reference block, as Figure 9 shown (the left side is the reference frame and the right side is the comparison frame). It includes the following steps:

[0072] First, in the comparison frame, define the search area of the starting tracking block in the transverse direction inside the blood vessel. The pixel size of the search area is M×N.

[0073] Secondly, define the matching blocks of the starting tracking block within the search area, and their pixel sizes are the same as those of the tracking blocks. Let the pixel coordinates of the matching block within the search area be (m, n), where 1 ≤ m ≤ M and 1 ≤ n ≤ N, that is, there are M × N matching blocks within the search area.

[0074] Then, according to formulas (1), (2), and (3), calculate the mean square differences (MSD) between all the matching blocks and the starting tracking block.

[0075]

[0076]

[0077]

[0078] Where x and y represent pixel values, X × Y is the pixel size of the starting tracking block and the corresponding matching block, X represents the pixel width, Y represents the pixel height, S(x, y) and T(x, y) respectively represent the pixel value matrices of the starting tracking block and the corresponding matching block, or E(S(x, y)) both represent the average pixel value of the starting tracking block, or E(T(x, y)) both represent the average pixel value of the corresponding matching block of the starting tracking block.

[0079] After that, according to formula (4), store the mean square difference values of the M × N matching blocks into the matrix of mean square differences (MMSD) at the corresponding positions.

[0080] MMSD(m, n) = MSD (m,n) (4)

[0081] Finally, find the minimum value in the mean square difference matrix MMSD, and record its pixel coordinates in the entire comparison frame as (K i , L i ). The matching block corresponding to the minimum value coordinates is the best matching block of the starting tracking block, and this matching block is defined as the reference block.

[0082] After step S3, as Figure 10 shown (the left side is the reference frame and the right side is the comparison frame), using the coordinates of the reference block as the starting point for the search, search for the best matching blocks of the remaining W - 1 tracking blocks, and record their pixel coordinates in the comparison frame. Among them, for the search areas of the remaining W - 1 tracking blocks, no fixed positions and sizes are set, but the positions of the search areas are determined by the pixel coordinates of the best matching blocks of adjacent tracking blocks, and the sizes of the search areas are determined by the minimum value of the mean square difference function.

[0083] Step S4: Taking the best matching block of the starting tracking block as a reference, in the order from bottom to top, sequentially determine the best matching blocks of all the tracking blocks above the starting tracking block in the comparison frame.

[0084] Step S4 specifically includes:

[0085] Step S4.1: Determine the best matching block of the starting tracking block as the first reference block, and determine the adjacent tracking block above the starting tracking block as the first current tracking block.

[0086] Step S4.2: Determine the adjacent matching block above the first reference block as the first intermediate matching block.

[0087] Step S4.3: Taking the first intermediate matching block as a search starting point, adopt a search strategy with an adaptive search area to determine the best matching block of the first current tracking block among several matching blocks on the left and right of the search starting point; the sum of mean squared errors between the best matching block of the first current tracking block and the first current tracking block is simultaneously less than the sum of mean squared errors between the matching block at one pixel to the left of the best matching block of the first current tracking block and the first current tracking block and the sum of mean squared errors between the matching block at one pixel to the right of the best matching block of the first current tracking block and the first current tracking block.

[0088] Step S4.4: Determine whether the first current tracking block is the last tracking block above the starting tracking block, and obtain a first judgment result.

[0089] Step S4.5: If the first judgment result is no, update the first reference block with the best matching block of the first current tracking block, update the first current tracking block with the adjacent tracking block above the first current tracking block, and return to Step S4.2.

[0090] Step S4.6: If the first judgment result is yes, the best matching blocks of all the tracking blocks above the starting tracking block are determined.

[0091] Specifically, sequentially search for the best matching blocks of the (i - 1)-th, (i - 2)-th, …, 1-st tracking blocks.

[0092] 1) When searching for the best matching block of the (i - 1)-th tracking block:

[0093] ① Taking the pixel coordinates (K i , L i ) of the best matching block (i.e., the reference block) of the i-th tracking block as a reference, calculate the pixel coordinates (K i , L i - Y) of the search starting point.

[0094] ② At (K i -1, L i -Y), (K i , L i -Y), and (K i +1, L i -Y), take matching blocks at 3 pixel coordinates, calculate the sum of mean squared errors of the 3 matching blocks and the (i - 1)-th tracking block, and form a sum-of-mean-squared-errors function with 3 values.

[0095] ③ Determine whether the sum of mean squared errors corresponding to the matching block at the pixel coordinate (K i , L i -Y) is the minimum among the 3 sum-of-mean-squared-errors functions. If so, the matching block at the pixel coordinate (K i , L i -Y) is the best matching block of the (i - 1)-th tracking block, and record its pixel coordinates in the entire comparison frame as (K i-1 , L i-1 ).

[0096] Determine whether the sum of mean squared errors corresponding to the matching block at the pixel coordinate (K i -1, L i -Y) is the minimum among the 3 sum-of-mean-squared-errors functions. If so, take a matching block at the pixel coordinate (K i -2, L i -Y), calculate the sum of mean squared errors of this matching block and the (i - 1)-th tracking block, and determine whether the MSD corresponding to the matching block at the pixel coordinate (K i -2, L i -Y) < the MSD corresponding to the matching block at the pixel coordinate (K i -1, L i -Y). If not, the matching block at the pixel coordinate (K i -1, L i -Y) is the best matching block of the (i - 1)-th tracking block, and record its pixel coordinates in the entire comparison frame as (K i-1 , L i-1 ). If so, repeat the steps, continue to shift the matching block to the left until it is found that the MSD corresponding to the matching block at the pixel coordinate (K i -h, L i -Y) > the MSD corresponding to the matching block at the pixel coordinate (K i -(h - 1), L i -Y), then the pixel coordinate (K i -(h - 1), L iThe matching block at (-Y) is the best matching block of the (i - 1)-th tracking block, and record its pixel coordinates in the entire comparison frame as (K i-1 , L i-1 ).

[0097] Judge whether the sum of mean squared errors corresponding to the matching block at pixel coordinates (K i + 1, L i - Y) is the minimum among the three sum of mean squared error functions. If so, take the matching block at pixel coordinates (K i + 2, L i - Y), calculate the sum of mean squared errors between this matching block and the (i - 1)-th tracking block, and judge whether the MSD corresponding to the matching block at pixel coordinates (K i + 2, L i - Y) < the MSD corresponding to the matching block at pixel coordinates (K i + 1, L i - Y). If not, then the matching block at pixel coordinates (K i + 1, L i - Y) is the best matching block of the (i - 1)-th tracking block, and record its pixel coordinates in the entire comparison frame as (K i-1 , L i-1 ). If so, repeat the steps, continue to shift the matching block to the right until it is found that the MSD corresponding to the matching block at pixel coordinates (K i + h, L i - Y) > the MSD corresponding to the matching block at pixel coordinates (K i + (h - 1), L i - Y), then the matching block at pixel coordinates (K i + (h - 1), L i - Y) is the best matching block of the (i - 1)-th tracking block, and record its pixel coordinates in the entire comparison frame as (K i-1 , L i-1 ).

[0098] 2) When searching for the best matching block of the (i - 2)-th tracking block:

[0099] ① Taking the pixel coordinates (K i-1 , L i-1 ) of the best matching block of the (i - 1)-th tracking block as a reference, calculate the pixel coordinates (K i-1 , L i-1 - Y) of the search starting point.

[0100] ② At (K i-1 - 1, L i-1 - Y), (K i-1 , L i-1 - Y), (K i-1+1, L i-1 Take matching blocks at 3 pixel coordinates of (-Y), calculate the sum of mean squared errors of the 3 matching blocks and the (i - 2)-th tracking block, and form a sum-of-mean-squared-errors function with 3 values.

[0101] ③ Judge (K i-1 , L i-1 Whether the sum of mean squared errors corresponding to the matching block at the (-Y) pixel coordinate is the minimum among the 3 sum-of-mean-squared-errors functions. If so, then (K i-1 , L i-1 The matching block at the (-Y) pixel coordinate is the best matching block of the (i - 2)-th tracking block, and record its pixel coordinates in the entire comparison frame as (K i-2 , L i-2 ).

[0102] Judge whether the sum of mean squared errors corresponding to the matching block at the (K i-1 -1, L i-1 -Y) pixel coordinate is the minimum among the 3 sum-of-mean-squared-errors functions. If so, then take a matching block at the (K i-1 -2, L i-1 -Y) pixel coordinate, calculate the sum of mean squared errors of this matching block and the (i - 2)-th tracking block, and judge whether the MSD corresponding to the matching block at the pixel coordinate (K i-1 -2, L i-1 -Y) < the MSD corresponding to the matching block at the pixel coordinate (K i-1 -1, L i-1 -Y). If not, then the matching block at the pixel coordinate (K i-1 -1, L i-1 -Y) is the best matching block of the (i - 2)-th tracking block, and record its pixel coordinates in the entire comparison frame as (K i-2 , L i-2 ). If so, repeat the steps, continue to shift the matching block to the left until it is found that the MSD corresponding to the matching block at the pixel coordinate (K i-1 -h, L i-1 -Y) > the MSD corresponding to the matching block at the pixel coordinate (K i-1 -(h - 1), L i-1 -Y). Then the matching block at the pixel coordinate (K i-1 -(h - 1), L i-1 -Y) is the best matching block of the (i - 2)-th tracking block, and record its pixel coordinates in the entire comparison frame as (K i-2 , L i-2 ).

[0103] Judge (K i-1 +1, L i-1- Whether the sum of mean squared differences (MSD) corresponding to the matching block at the pixel coordinates (K i-1 +2, L i-1 -Y) is the minimum among the three MSD functions. If so, take the matching block at the pixel coordinates (K i-1 +2, L i-1 -Y), calculate the MSD between this matching block and the (i - 2)-th tracked block, and determine whether the MSD of the matching block at the pixel coordinates (K i-1 +1, L i-1 -Y) < the MSD of the matching block at the pixel coordinates (K i-1 +1, L i-1 -Y). If not, the matching block at the pixel coordinates (K i-2 , L i-2 ) is the best matching block of the (i - 2)-th tracked block, and record its pixel coordinates in the entire comparison frame as (K i-1 +h, L i-1 -Y). If so, repeat the steps and continue to shift the matching block to the right until it is found that the MSD of the matching block at the pixel coordinates (K i-1 +(h - 1), L i-1 -Y) > the MSD of the matching block at the pixel coordinates (K i-1 +(h - 1), L i-1 -Y). Then, the matching block at the pixel coordinates (K i-2 , L i-2 ) is the best matching block of the (i - 2)-th tracked block, and record its pixel coordinates in the entire comparison frame as (K

[0104] 3) As described above, when searching for the best matching block of the (i - n)-th tracked block, taking the pixel coordinates of the best matching block of the (i - n + 1)-th tracked block as a reference, repeat the search steps until the search for the best matching block of the 1st tracked block is completed. Such a search method can not only save time but also eliminate the abnormal MSD interference caused by speckle decorrelation between images.

[0105] Step S5: Taking the best matching block of the starting tracked block as a reference, in the order from top to bottom, sequentially determine the best matching blocks of all the tracked blocks below the starting tracked block in the comparison frame.

[0106] Step S5 specifically includes:

[0107] Step S5.1: Determine the best matching block of the starting tracked block as the second reference block, and determine the adjacent tracked block below the starting tracked block as the second current tracked block.

[0108] Step S5.2: Determine the adjacent matching block below the second reference block as the second intermediate matching block.

[0109] Step S5.3: Taking the second intermediate matching block as the search starting point, adopting a search strategy with an adaptive search area, determine the best matching block of the second current tracking block among several matching blocks on the left and right of the search starting point; the sum of the mean squared errors between the best matching block of the second current tracking block and the second current tracking block is simultaneously less than the sum of the mean squared errors between the matching block at one pixel to the left of the best matching block of the second current tracking block and the second current tracking block, and the sum of the mean squared errors between the matching block at one pixel to the right of the best matching block of the second current tracking block and the second current tracking block.

[0110] Step S5.4: Determine whether the second current tracking block is the last tracking block below the starting tracking block, obtaining a second judgment result.

[0111] Step S5.5: If the second judgment result is no, update the second reference block with the best matching block of the second current tracking block, update the second current tracking block with the adjacent tracking block below the second current tracking block, and return to Step S5.2.

[0112] Step S5.6: If the second judgment result is yes, the best matching blocks of all tracking blocks below the starting tracking block are determined.

[0113] Specifically, sequentially search for the best matching blocks of the (i + 1)-th, (i + 2)-th, …, W-th tracking blocks.

[0114] 1) When searching for the best matching block of the (i + 1)-th tracking block:

[0115] ① Taking the pixel coordinates (K i , L i ) of the best matching block of the i-th tracking block (i.e., the reference block) as a reference, calculate the pixel coordinates (K i , L i + Y) of the search starting point.

[0116] ② Take matching blocks at the three pixel coordinates (K i - 1, L i + Y), (K i , L i + Y), (K i + 1, L i + Y), calculate the sum of the mean squared errors between the three matching blocks and the (i + 1)-th tracking block, and form a sum-of-mean-squared-errors function with three values.

[0117] ③ Determine (K i , L iWhether the sum of mean squared errors (MSD) corresponding to the matching block at the pixel coordinates of (K i , L i + Y) is the minimum among the three MSD functions. If so, then (K i , L i + Y) is the best matching block of the (i + 1)-th tracking block, and record its pixel coordinates in the entire comparison frame as (K i+1 , L i+1 ).

[0118] Judge whether the sum of mean squared errors (MSD) corresponding to the matching block at the pixel coordinates of (K i - 1, L i + Y) is the minimum among the three MSD functions. If so, then take the matching block at the pixel coordinates of (K i - 2, L i + Y), calculate the MSD between this matching block and the (i + 1)-th tracking block, and judge whether the MSD of the matching block at the pixel coordinates of (K i - 2, L i + Y) < the MSD of the matching block at the pixel coordinates of (K i - 1, L i + Y). If not, then the matching block at the pixel coordinates of (K i - 1, L i + Y) is the best matching block of the (i + 1)-th tracking block, and record its pixel coordinates in the entire comparison frame as (K i+1 , L i+1 ). If so, repeat the steps and continue to shift the matching block to the left until it is found that the MSD of the matching block at the pixel coordinates of (K i - h, L i + Y) > the MSD of the matching block at the pixel coordinates of (K i - (h - 1), L i + Y). Then the matching block at the pixel coordinates of (K i - (h - 1), L i + Y) is the best matching block of the (i + 1)-th tracking block, and record its pixel coordinates in the entire comparison frame as (K i+1 , L i+1 ).

[0119] Judge whether the sum of mean squared errors (MSD) corresponding to the matching block at the pixel coordinates of (K i + 1, L i + Y) is the minimum among the three MSD functions. If so, then take the matching block at the pixel coordinates of (K i + 2, L i + Y), calculate the MSD between this matching block and the (i + 1)-th tracking block, and judge whether the pixel coordinates of (K i + 2, Li The MSD corresponding to the matching block on (+Y), and the pixel coordinates (K i +1, L i The MSD corresponding to the matching block on (+Y). If not, then the pixel coordinates (K i +1, L i The matching block on (+Y) is the best matching block of the (i + 1)-th tracking block, and record its pixel coordinates in the entire comparison frame as (K i+1 , L i+1 ). If so, repeat the steps and continue to shift the matching block to the right until finding the pixel coordinates (K i +h, L i The MSD corresponding to the matching block on (+Y) > the pixel coordinates (K i +(h - 1), L i The MSD corresponding to the matching block on (+Y), then the pixel coordinates (K i +(h - 1), L i The matching block on (+Y) is the best matching block of the (i + 1)-th tracking block, and record its pixel coordinates in the entire comparison frame as (K i+1 , L i+1 ).

[0120] 2) When searching for the best matching block of the (i + 2)-th tracking block:

[0121] ① Using the pixel coordinates (K i+1 , L i+1 ) of the best matching block of the (i + 1)-th tracking block as a reference, calculate the pixel coordinates (K i+1 , L i+1 +Y) of the search starting point.

[0122] ② Take the matching blocks at the 3 pixel coordinates of (K i+1 -1, L i+1 +Y), (K i+1 , L i+1 +Y), (K i+1 +1, L i+1 +Y), calculate the sum of the mean squared errors of the 3 matching blocks and the (i + 2)-th tracking block, and form a mean squared error sum function with 3 values.

[0123] ③ Determine whether the sum of the mean squared errors corresponding to the matching block at the pixel coordinates (K i+1 , L i+1 +Y) is the minimum value among the 3 mean squared error sum functions. If so, then the matching block at the pixel coordinates (K i+1 , L i+1 +Y) is the best matching block of the (i + 2)-th tracking block, and record its pixel coordinates in the entire comparison frame as (K i+2 , Li+2 )。

[0124] Determine whether the sum of mean squared differences (MSD) corresponding to the matching block at the pixel coordinates (K i+1 -1, L i+1 +Y) is the minimum among the MSDs of three functions. If so, take the matching block at the pixel coordinates (K i+1 -2, L i+1 +Y), calculate the MSD between this matching block and the (i + 2)-th tracking block, and determine whether the MSD corresponding to the matching block at the pixel coordinates (K i+1 -2, L i+1 +Y) < the MSD corresponding to the matching block at the pixel coordinates (K i+1 -1, L i+1 +Y). If not, the matching block at the pixel coordinates (K i+1 -1, L i+1 +Y) is the best matching block for the (i + 2)-th tracking block, and record its pixel coordinates in the entire comparison frame as (K i+2 , L i+2 ). If so, repeat the steps and continue to shift the matching block to the left until it is found that the MSD corresponding to the matching block at the pixel coordinates (K i+1 -h, L i+1 +Y) > the MSD corresponding to the matching block at the pixel coordinates (K i+1 -(h - 1), L i+1 +Y). Then, the matching block at the pixel coordinates (K i+1 -(h - 1), L i+1 +Y) is the best matching block for the (i + 2)-th tracking block, and record its pixel coordinates in the entire comparison frame as (K i+2 , L i+2 ).

[0125] Determine whether the sum of mean squared differences (MSD) corresponding to the matching block at the pixel coordinates (K i+1 +1, L i+1 +Y) is the minimum among the MSDs of three functions. If so, take the matching block at the pixel coordinates (K i+1 +2, L i+1 +Y), calculate the MSD between this matching block and the (i + 2)-th tracking block, and determine whether the MSD corresponding to the matching block at the pixel coordinates (K i+1 +2, L i+1 +Y) < the MSD corresponding to the matching block at the pixel coordinates (K i+1 +1, L i+1 +Y). If not, the pixel coordinates (K i-1 +1, L i-1The matching block on (+Y) is the best matching block of the (i + 2)-th tracking block, and record its pixel coordinates in the entire comparison frame as (K i+2 , L i+2 ). If so, repeat the steps and continue to shift the matching block to the right until finding that the MSD corresponding to the matching block on the pixel coordinates (K i+1 +h, L i+1 +Y) > the MSD corresponding to the matching block on the pixel coordinates (K i+1 +(h - 1), L i+1 +Y). Then, the matching block on the pixel coordinates (K i+1 +(h - 1), L i+1 +Y) is the best matching block of the (i + 2)-th tracking block, and record its pixel coordinates in the entire comparison frame as (K i+2 , L i+2 ).

[0126] 3) As described above, when searching for the best matching block of the (i + n)-th tracking block, with the pixel coordinates of the best matching block of the (i + n - 1)-th tracking block as a reference, repeat the search steps until completing the search for the best matching block of the W-th tracking block. Such a search method can not only save time but also eliminate the abnormal MSD interference caused by speckle decorrelation between images.

[0127] Step S6: Determine the blood flow velocity corresponding to each tracking block according to the pixel coordinates of all tracking blocks and the pixel coordinates of all best matching blocks respectively.

[0128] Step S6 specifically includes: determining the displacement of each tracking block according to the pixel coordinates of all tracking blocks and the pixel coordinates of all best matching blocks respectively; determining the blood flow velocity corresponding to each tracking block according to the displacements of all tracking blocks and the acquisition duration between the reference frame and the comparison frame respectively.

[0129] Specifically, according to the search results of Step S4 and Step S5, record the pixel coordinates (K w , L w ) of the best matching blocks of the W tracking blocks, where 1 ≤ w ≤ W.

[0130] Calculate the displacement D of the W tracking blocks using formula (5),

[0131]

[0132] where l pixel represents the physical length of one pixel.

[0133] Calculate the blood flow velocity V corresponding to the W tracking blocks using formula (6),

[0134]

[0135] Among them, FR represents the acquisition duration between the reference frame and the comparison frame.

[0136] Furthermore, it also includes: determining a blood flow parabola according to the blood flow velocities corresponding to all tracking blocks. As Figure 11 shown, connecting W blood flow velocity measurement values can obtain a blood flow parabola.

[0137] A specific embodiment is provided below to elaborate on the above method in detail.

[0138] 1. Algorithm initialization. It includes the following steps:

[0139] 1.1 Use ultrasonic scanning to generate a time series of B-mode ultrasound images of the vascular radial cross-section.

[0140] 1.2 Randomly select one frame of the image in the time series of ultrasonic images of the vascular radial cross-section as the reference frame, and use the subsequent frame of the reference frame as the comparison frame.

[0141] 1.3 Along the position of the radial blood flow in the reference frame, select W = 30 windows with a pixel size of X×Y = 30×18 as tracking blocks, as Figure 8 shown. Denote the pixel coordinates of these 30 tracking blocks in the reference frame as (H w , T w ), where 1 ≤ w ≤ W.

[0142] 1.4 Among the 30 tracking blocks from the upper blood vessel wall to the lower blood vessel wall, select the tracking blocks with serial numbers between 0.4W and 0.6W = 12 and 18 as the starting tracking blocks. The position of the starting tracking blocks is as Figure 8 shown, and the serial number of the starting tracking block here is 12.

[0143] 2. In the search area of the starting tracking block, find the best matching block of the starting tracking block as the reference block, as Figure 9 shown. It includes the following steps:

[0144] 2.1 In the comparison frame, define the search area of the starting tracking block along the internal horizontal direction of the blood vessel. The pixel size of the search area is M×N = 50×41.

[0145] 2.2 Define the matching block of the starting tracking block in the search area, and its pixel size is the same as that of the tracking block. Let the pixel coordinates of the matching block in the search area be (m, n), where 1 ≤ m ≤ M and 1 ≤ n ≤ N, that is, there are M×N = 50×41 matching blocks in the search area.

[0146] 2.3 According to formulas (1), (2), and (3), calculate the sum of the mean squared errors of all matching blocks and the starting tracking block.

[0147] 2.4 According to formula (4), store the mean squared error values of M×N = 50×41 matching blocks into the mean squared error matrix at the corresponding positions.

[0148] 2.5 Search for the minimum value in the mean squared error matrix MMSD, and record its pixel coordinates in the entire comparison frame as (K 12 , L 12 ). The matching block corresponding to the minimum value coordinates is the best matching block of the starting tracking block, and this matching block is defined as the reference block.

[0149] 3. As Figure 10 shown, taking the coordinates of the reference block as the starting point of the search, search for the best matching blocks of the remaining W - 1 = 29 tracking blocks, and record their pixel coordinates in the comparison frame. Among them, for the search areas of the remaining 29 tracking blocks, the fixed positions and sizes are not set anymore. Instead, the position of the search area is determined by the pixel coordinates of the best matching blocks of adjacent tracking blocks, and the size of the search area is determined by the minimum value of the mean squared error function.

[0150] 3.1 Search for the best matching blocks of the 11th, 10th, …, 1st tracking blocks in sequence.

[0151] 3.1.1 When searching for the best matching block of the 11th tracking block: ① Taking the pixel coordinates (K 12 , L 12 ) of the best matching block of the 12th tracking block (i.e., the reference block) as a reference, calculate the pixel coordinates (K 12 , L 12 - 18) of the search starting point. ② Take matching blocks at the 3 pixel coordinates of (K 12 - 1, L 12 - 18), (K 12 , L 12 - 18), (K 12 + 1, L 12 - 18), calculate the mean squared error between the 3 matching blocks and the 11th tracking block, and form a mean squared error function with 3 values. ③ Determine whether the mean squared error corresponding to the matching block at the pixel coordinates (K 12 , L 12 - 18) is the minimum value among the 3 mean squared error functions. If so, the matching block at the pixel coordinates (K 12 , L 12 - 18) is the best matching block of the 11th tracking block, and record its pixel coordinates in the entire comparison frame as (K 11 , L 11 ). Determine whether the mean squared error corresponding to the matching block at the pixel coordinates (K 12 - 1, L 12-12) Whether the sum of mean squared differences (MSD) corresponding to the matching block at the pixel coordinates is the minimum among the three MSD functions. If so, at (K 12 -2, L 12 -18), take the matching block at the pixel coordinates, calculate the MSD between this matching block and the 11th tracking block, and determine whether the MSD of the matching block at the pixel coordinates (K 12 -2, L 12 -Y) < the MSD of the matching block at the pixel coordinates (K 12 -1, L 12 -18). If not, the matching block at the pixel coordinates (K 12 -1, L 12 -18) is the best matching block of the 11th tracking block, and record its pixel coordinates in the entire comparison frame as (K 11 , L 11 ). If so, repeat the steps, continue to shift the matching block to the left until it is found that the MSD of the matching block at the pixel coordinates (K 12 -h, L 12 -18) > the MSD of the matching block at the pixel coordinates (K 12 -(h - 1), L 12 -18). Then, the matching block at the pixel coordinates (K 12 -(h - 1), L 12 -18) is the best matching block of the 11th tracking block, and record its pixel coordinates in the entire comparison frame as (K 11 , L 11 ). Determine whether the sum of mean squared differences corresponding to the matching block at the pixel coordinates (K 12 +1, L 12 -18) is the minimum among the three sum of mean squared differences functions. If so, at (K 12 +2, L 12 -18), take the matching block at the pixel coordinates, calculate the MSD between this matching block and the 11th tracking block, and determine whether the MSD of the matching block at the pixel coordinates (K 12 +2, L 12 -18) < the MSD of the matching block at the pixel coordinates (K 12 +1, L 12 -18). If not, the matching block at the pixel coordinates (K 12 +1, L 12 -18) is the best matching block of the 11th tracking block, and record its pixel coordinates in the entire comparison frame as (K 11 , L 11 ). If so, repeat the steps, continue to shift the matching block to the right until it is found that the MSD of the matching block at the pixel coordinates (K 12 +h, L12 -18), the MSD corresponding to the matching block at pixel coordinates (K 12 +(h - 1), L 12 -18), if the MSD corresponding to the matching block at pixel coordinates (K 12 +(h - 1), L 12 -18) is the best matching block of the 11th tracking block, record its pixel coordinates in the entire comparison frame as (K 11 , L 11 ).

[0152] 3.1.2 When searching for the best matching block of the 10th tracking block: ① Using the pixel coordinates (K 11 , L 11 ) of the best matching block of the 11th tracking block as a reference, calculate the pixel coordinates (K 11 , L 11 -18) of the search starting point. ② Take matching blocks at the 3 pixel coordinates of (K 11 -1, L 11 -18), (K 11 , L 11 -18), and (K 11 +1, L 11 -18), calculate the sum of the mean squared errors of the 3 matching blocks and the 10th tracking block, and form a mean squared error sum function with 3 values. ③ Determine whether the sum of the mean squared errors corresponding to the matching block at the pixel coordinates (K 11 , L 11 -18) is the minimum value among the 3 mean squared error sum functions. If so, the matching block at the pixel coordinates (K 11 , L 11 -18) is the best matching block of the 10th tracking block, and record its pixel coordinates in the entire comparison frame as (K 10 , L 10 ). Determine whether the sum of the mean squared errors corresponding to the matching block at the pixel coordinates (K 11 -1, L 11 -18) is the minimum value among the 3 mean squared error sum functions. If so, take a matching block at the pixel coordinates (K 11 -2, L 11 -18), calculate the sum of the mean squared errors of this matching block and the 10th tracking block, and determine whether the MSD corresponding to the matching block at the pixel coordinates (K 11 -2, L 11 -18) < the MSD corresponding to the matching block at the pixel coordinates (K 11 -1, L 11 -18). If not, the pixel coordinates (K 11 -1, L 11-18) is the best matching block for the 10th tracking block, and record its pixel coordinates in the entire comparison frame as (K 10 , L 10 ). If so, repeat the step and continue to shift the matching block to the left until finding that the MSD of the matching block at pixel coordinates (K 11 -h, L 11 -18) > the MSD of the matching block at pixel coordinates (K 11 -(h - 1), L 11 -18), then the matching block at pixel coordinates (K 11 -(h - 1), L 11 -18) is the best matching block for the 10th tracking block, and record its pixel coordinates in the entire comparison frame as (K 10 , L 10 ). Determine whether the sum of mean squared differences of the matching block at pixel coordinates (K 11 +1, L 11 -18) is the minimum value among the three sum of mean squared difference functions. If so, take the matching block at pixel coordinates (K 11 +2, L 11 -18), calculate the sum of mean squared differences between this matching block and the 10th tracking block, and determine whether the MSD of the matching block at pixel coordinates (K 11 +2, L 11 -18) < the MSD of the matching block at pixel coordinates (K 11 +1, L 11 -18). If not, then the matching block at pixel coordinates (K 11 +1, L 11 -Y) is the best matching block for the 10th tracking block, and record its pixel coordinates in the entire comparison frame as (K 10 , L 10 ). If so, repeat the step and continue to shift the matching block to the right until finding that the MSD of the matching block at pixel coordinates (K 11 +h, L 11 -18) > the MSD of the matching block at pixel coordinates (K 11 +(h - 1), L 11 -18), then the matching block at pixel coordinates (K 11 +(h - 1), L 11 -18) is the best matching block for the 10th tracking block, and record its pixel coordinates in the entire comparison frame as (K 11 , L 11 ).

[0153] 3.1.3 As described in 3.1.1 and 3.1.2, when searching for the best matching block of the 12 - nth tracking block, the pixel coordinates of the best matching block of the 13 - nth tracking block are used as a reference, and the search steps are repeated until the search for the best matching block of the 1st tracking block is completed.

[0154] 3.2 Search for the best matching blocks of the 13th, 14th, …, 30th tracking blocks in sequence.

[0155] 3.2.1 When searching for the best matching block of the 13th tracking block: ① Using the pixel coordinates (K 12 , L 12 ) of the best matching block of the 12th tracking block (i.e., the reference block) as a reference, calculate the pixel coordinates (K 12 , L 12 + 18) of the search starting point. ② Take matching blocks at the three pixel coordinates of (K 12 - 1, L 12 + 18), (K 12 , L 12 + 18), and (K 12 + 1, L 12 + 18), calculate the sum of the squared mean errors of the three matching blocks and the 13th tracking block, and form a sum - of - squared - mean - error function with three values. ③ Determine whether the sum of the squared mean errors corresponding to the matching block at the pixel coordinate (K 12 , L 12 + 18) is the minimum value among the three sum - of - squared - mean - error functions. If so, the matching block at the pixel coordinate (K 12 , L 12 + 18) is the best matching block of the 13th tracking block, and record its pixel coordinates in the entire comparison frame as (K 13 , L 13 ). Determine whether the sum of the squared mean errors corresponding to the matching block at the pixel coordinate (K 12 - 1, L 12 + 18) is the minimum value among the three sum - of - squared - mean - error functions. If so, take a matching block at the pixel coordinate (K 12 - 2, L 12 + 18), calculate the sum of the squared mean errors of this matching block and the 13th tracking block, and determine whether the MSD of the matching block at the pixel coordinate (K 12 - 2, L 12 + 18) < the MSD of the matching block at the pixel coordinate (K 12 - 1, L 12 + 18). If not, the matching block at the pixel coordinate (K 12 - 1, L 12 + 18) is the best matching block of the 13th tracking block, and record its pixel coordinates in the entire comparison frame as (K13 , L 13 ). If so, repeat the steps and continue to shift the matching block to the left until the MSD corresponding to the matching block at the pixel coordinates (K 12 -h, L 12 +18) is greater than the MSD corresponding to the matching block at the pixel coordinates (K 12 -(h - 1), L 12 +18). Then, the matching block at the pixel coordinates (K 12 -(h - 1), L 12 +18) is the best matching block of the 13th tracking block, and record its pixel coordinates in the entire comparison frame as (K 13 , L 13 ). Determine whether the sum of the squared mean errors corresponding to the matching block at the pixel coordinates (K 12 +1, L 12 +18) is the minimum among the three sum of squared mean error functions. If so, take the matching block at the pixel coordinates (K 12 +2, L 12 +18), calculate the sum of the squared mean errors between this matching block and the 13th tracking block, and determine whether the MSD corresponding to the matching block at the pixel coordinates (K 12 +2, L 12 +18) is less than the MSD corresponding to the matching block at the pixel coordinates (K 12 +1, L 12 +18). If not, then the matching block at the pixel coordinates (K 12 +1, L 12 +18) is the best matching block of the 13th tracking block, and record its pixel coordinates in the entire comparison frame as (K 13 , L 13 ). If so, repeat the steps and continue to shift the matching block to the right until the MSD corresponding to the matching block at the pixel coordinates (K 12 +h, L 12 +18) is greater than the MSD corresponding to the matching block at the pixel coordinates (K 12 +(h - 1), L 12 +18). Then, the matching block at the pixel coordinates (K 12 +(h - 1), L 12 +18) is the best matching block of the 13th tracking block, and record its pixel coordinates in the entire comparison frame as (K 13 , L 13 ).

[0156] 3.1.2 When searching for the best matching block of the 14th tracking block: ① Use the pixel coordinates (K 13 , L 13)For reference, calculate the pixel coordinates (K 13 , L 13 + 18). ② Take matching blocks at the 3 pixel coordinates of (K 13 - 1, L 13 + 18), (K 13 , L 13 + 18), (K 13 + 1, L 13 + 18), calculate the sum of mean squared errors of the 3 matching blocks and the 14th tracking block, and form a sum-of-mean-squared-errors function with 3 values. ③ Determine whether the sum of mean squared errors corresponding to the matching block at the pixel coordinate of (K 13 , L 13 + 18) is the minimum among the 3 sum-of-mean-squared-errors functions. If so, the matching block at the pixel coordinate of (K 13 , L 13 + 18) is the best matching block of the 14th tracking block, and record its pixel coordinates in the entire comparison frame as (K 14 , L 14 ). Determine whether the sum of mean squared errors corresponding to the matching block at the pixel coordinate of (K 13 - 1, L 13 + 18) is the minimum among the 3 sum-of-mean-squared-errors functions. If so, take a matching block at the pixel coordinate of (K 13 - 2, L 13 + 18), calculate the sum of mean squared errors of this matching block and the 14th tracking block, and determine whether the MSD corresponding to the matching block at the pixel coordinate of (K 13 - 2, L 13 + 18) < the MSD corresponding to the matching block at the pixel coordinate of (K 13 - 1, L 13 + 18). If not, the matching block at the pixel coordinate of (K 13 - 1, L 13 + 18) is the best matching block of the 14th tracking block, and record its pixel coordinates in the entire comparison frame as (K 14 , L 14 ). If so, repeat the steps and continue to shift the matching block to the left until it is found that the MSD corresponding to the matching block at the pixel coordinate of (K 13 - h, L 13 + 18) > the MSD corresponding to the matching block at the pixel coordinate of (K 13 - (h - 1), L 13 + 18), then the matching block at the pixel coordinate of (K 13 - (h - 1), L 13 + 18) is the best matching block of the 14th tracking block, and record its pixel coordinates in the entire comparison frame as (K14 , L 14 ). Determine if the sum of squared mean errors corresponding to the matching block at pixel coordinates (K 13 + 1, L 13 + 18) is the minimum among the three sum of squared mean error functions. If so, take the matching block at pixel coordinates (K 13 + 2, L 13 + 18), calculate the sum of squared mean errors between this matching block and the 14th tracking block, and determine if the MSD corresponding to the matching block at pixel coordinates (K 13 + 2, L 13 + 18) < the MSD corresponding to the matching block at pixel coordinates (K 13 + 1, L 13 + 18). If not, the matching block at pixel coordinates (K 13 + 1, L 13 + 18) is the best matching block of the 14th tracking block, and record its pixel coordinates in the entire comparison frame as (K 14 , L 14 ). If so, repeat the steps, continue to shift the matching block to the right until it is found that the MSD corresponding to the matching block at pixel coordinates (K 13 + h, L 13 + 18) > the MSD corresponding to the matching block at pixel coordinates (K 13 + (h - 1), L 13 + 18). Then, the matching block at pixel coordinates (K 13 + (h - 1), L 13 + 18) is the best matching block of the 14th tracking block, and record its pixel coordinates in the entire comparison frame as (K 14 , L 14 ).

[0157] 3.2.3 As described in 3.2.1 and 3.2.2, when searching for the best matching block of the (12 + n)th tracking block, use the pixel coordinates of the best matching block of the (11 + n)th tracking block as a reference, and repeat the search steps until the search for the best matching block of the 30th tracking block is completed.

[0158] 3.3 According to the search results in 3.1 and 3.2, record the pixel coordinates (K w , L w ) of the best matching blocks of 30 tracking blocks.

[0159] 4. Calculate the displacement D of 30 tracking blocks according to formula (5), where l pixel = 7.1×10 -5 m represents the physical length of one pixel.

[0160] 5. Calculate the blood flow velocity V corresponding to 30 tracking blocks according to formula (6), where FR = 9×10 -4 s is the acquisition duration between the reference frame and the comparison frame. Connecting the 30 blood flow velocity measurement values can obtain a blood flow parabola, as Figure 12 shown.

[0161] Due to the certain viscosity of blood, the radial profile of blood flow velocity has the characteristics of laminar flow, which results in two adjacent tracking blocks, and their best matching blocks are not far away. Therefore, the position of the search area for each tracking block can be determined according to the position of the best matching block of the adjacent tracking block. The improvements of the present invention compared with the prior art are mainly two points. One is to change the search strategy, no longer perform traversal search on a large search area, but perform extended search around the position of the best matching block of the adjacent tracking block, and stop searching once the minimum value of the mean squared error is determined; the other is to consider the correlation between two adjacent tracking blocks, and the position of the search area for each tracking block is determined according to the position of the best matching block of the adjacent tracking block.

[0162] The above improvements bring three advantages. One is that it no longer performs traversal search on a large search area, saving the measurement time of blood flow velocity; the second is that it no longer performs independent search on each tracking block, considering the velocity correlation between each tracking block, avoiding local outliers, and improving the measurement accuracy of blood flow velocity; the third is that the range of the search area is not limited in size, and the search range can be infinitely expanded before finding the minimum value of the mean squared error, so that the speckle tracking method is no longer limited by the maximum detectable velocity, expanding the measurement range of blood flow velocity.

[0163] Embodiment 2

[0164] In order to execute the method corresponding to the above Embodiment 1 to achieve the corresponding functions and technical effects, the following provides an ultrasonic speckle tracking blood flow velocity measurement system, including:

[0165] An image acquisition module, configured to acquire an ultrasonic image sequence of the radial section of a blood vessel, and determine a reference frame and a comparison frame from the ultrasonic image sequence; the comparison frame is the subsequent frame image of the reference frame.

[0166] A tracking block determination module, configured to determine a plurality of tracking blocks along the position where the radial blood flow is located in the reference frame, and determine any one of the tracking blocks in the middle area as the starting tracking block.

[0167] A best matching block determination module for the starting tracking block, configured to determine the search area of the starting tracking block along the transverse direction inside the blood vessel in the comparison frame, and determine the best matching block of the starting tracking block within the search area of the starting tracking block.

[0168] The best matching block determination module for the upper tracking block is configured to, taking the best matching block of the starting tracking block as a reference, determine the best matching blocks of all the tracking blocks above the starting tracking block in the comparison frame in sequence from bottom to top.

[0169] The best matching block determination module for the lower tracking block is configured to, taking the best matching block of the starting tracking block as a reference, determine the best matching blocks of all the tracking blocks below the starting tracking block in the comparison frame in sequence from top to bottom.

[0170] The blood flow velocity determination module is configured to determine the blood flow velocity corresponding to each tracking block respectively according to the pixel coordinates of all the tracking blocks and the pixel coordinates of all the best matching blocks.

[0171] Embodiment III

[0172] The embodiment of the present invention further provides an electronic device, including a memory and a processor. The memory is used to store a computer program, and the processor is used to run the computer program so that the electronic device executes the ultrasonic speckle tracking blood flow velocity measurement method in Embodiment I. The electronic device may be a server.

[0173] In addition, the present invention further provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the ultrasonic speckle tracking blood flow velocity measurement method in Embodiment I is implemented.

[0174] In summary, in view of the disadvantages of the four existing technologies mentioned in the background art, the present invention proposes an ultrasonic speckle tracking blood flow velocity measurement method, system, device and medium with an adaptive search area. This technology no longer uses a search area with the same size and fixed position, but based on the laminar flow characteristics of the blood flow velocity, adaptively updates the position of the search area in the next search process according to the previous search result, and adaptively updates the size of the search area according to the minimum value of the cross-correlation function (i.e., the mean square error function). The purpose of this improvement is to make the search process more accurate and fast, so that the measurement accuracy and efficiency of the blood flow velocity are better and not limited by the maximum detectable blood flow velocity.

[0175] The present invention mainly solves the problems of long algorithm running time and inaccurate measurement results caused by global traversal in the traditional speckle tracking blood flow velocity measurement method. Compared with the traditional ultrasonic speckle tracking blood flow velocity measurement method, the present invention has three advantages. First, it improves the efficiency of the algorithm for collecting blood flow velocity in ultrasonic speckle tracking blood flow velocity measurement, avoids calculation redundancy in the process, and can effectively avoid the situation of delaying the disease due to long velocity measurement time. Second, it improves the accuracy of the blood flow velocity measurement result, avoids the interference of abnormal values caused by speckle decorrelation in the ultrasonic speckle tracking algorithm, and thus reduces the situation of inaccurate blood flow velocity measurement. Third, it is not limited by the maximum detectable blood flow velocity, and solves the problem that the maximum detectable blood flow velocity of the traditional ultrasonic speckle tracking blood flow velocity measurement is limited by the fixed-size search area of two adjacent ultrasonic images.

[0176] The various embodiments in this specification are described in a progressive manner. The key point of each embodiment is to illustrate the differences from other embodiments. The same or similar parts among the various embodiments can be referred to each other. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the description in the method part.

[0177] Specific examples are used in this article to elaborate on the principle and implementation manner of the present invention. The description of the above embodiments is only used to help understand the method and its core idea of the present invention. At the same time, for those of ordinary skill in the art, based on the idea of the present invention, there will be changes in the specific implementation manner and application scope. In summary, the content of this specification should not be construed as a limitation to the present invention.

Claims

1. An ultrasonic speckle tracking blood flow velocity measurement method, characterized in that, Including: Obtaining an ultrasonic image sequence of a radial section of a blood vessel, and determining a reference frame and a comparison frame from the ultrasonic image sequence; The comparison frame is the subsequent frame image of the reference frame; Determining a plurality of tracking blocks along the position where the radial blood flow is located in the reference frame, and determining any one of the tracking blocks in the middle area as the starting tracking block; Determining a search area of the starting tracking block along the transverse direction inside the blood vessel in the comparison frame, and determining the best matching block of the starting tracking block within the search area of the starting tracking block; Taking the best matching block of the starting tracking block as a reference, in the order from bottom to top, successively determining the best matching blocks of all the tracking blocks above the starting tracking block in the comparison frame, specifically including: Determining the best matching block of the starting tracking block as the first reference block, and determining the adjacent tracking block above the starting tracking block as the first current tracking block; Determining the adjacent matching block above the first reference block as the first intermediate matching block; Taking the first intermediate matching block as a search starting point, adopting a search strategy with an adaptive search area, and determining the best matching block of the first current tracking block among several matching blocks on the left and right of the search starting point; the sum of the mean squared errors between the best matching block of the first current tracking block and the first current tracking block is simultaneously less than the sum of the mean squared errors between the matching block at a pixel point to the left of the best matching block of the first current tracking block and the first current tracking block and the sum of the mean squared errors between the matching block at a pixel point to the right of the best matching block of the first current tracking block and the first current tracking block; Judging whether the first current tracking block is the last tracking block above the starting tracking block, and obtaining a first judgment result; If the first judgment result is negative, updating the first reference block with the best matching block of the first current tracking block, updating the first current tracking block with the adjacent tracking block above the first current tracking block, and returning to the step of "determining the adjacent matching block above the first reference block as the first intermediate matching block"; If the first judgment result is positive, the best matching blocks of all the tracking blocks above the starting tracking block are determined; Taking the best matching block of the starting tracking block as a reference, in the order from top to bottom, successively determining the best matching blocks of all the tracking blocks below the starting tracking block in the comparison frame; Respectively determining the blood flow velocity corresponding to each tracking block according to the pixel coordinates of all the tracking blocks and the pixel coordinates of all the best matching blocks.

2. The ultrasonic speckle tracking blood flow velocity measurement method according to claim 1, characterized in that, Determining the search area of the starting tracking block along the transverse direction inside the blood vessel in the comparison frame, and determining the best matching block of the starting tracking block within the search area of the starting tracking block, specifically including: Determining the search area of the starting tracking block along the transverse direction inside the blood vessel in the comparison frame; Determining a plurality of matching blocks of the starting tracking block within the search area of the starting tracking block; the matching blocks of the starting tracking block have the same size as the starting tracking block; Calculating the sum of the mean squared errors between all the matching blocks of the starting tracking block and the starting tracking block; Determine the matching block of the starting tracking block with the minimum sum of mean squared errors as the best matching block of the starting tracking block.

3. The ultrasonic speckle tracking blood flow velocity measurement method according to claim 1, characterized in that Taking the best matching block of the starting tracking block as a reference, in the order from top to bottom, sequentially determine the best matching blocks of all tracking blocks below the starting tracking block in the comparison frame, specifically including: Determine the best matching block of the starting tracking block as the second reference block, and determine the adjacent tracking block below the starting tracking block as the second current tracking block; Determine the adjacent matching block below the second reference block as the second intermediate matching block; Taking the second intermediate matching block as the search starting point, adopt a search strategy with an adaptive search area to determine the best matching block of the second current tracking block among several matching blocks on the left and right of the search starting point; the sum of mean squared errors between the best matching block of the second current tracking block and the second current tracking block is simultaneously less than the sum of mean squared errors between the matching block at one pixel to the left of the best matching block of the second current tracking block and the second current tracking block, and the sum of mean squared errors between the matching block at one pixel to the right of the best matching block of the second current tracking block and the second current tracking block; Judge whether the second current tracking block is the last tracking block below the starting tracking block to obtain a second judgment result; If the second judgment result is no, update the second reference block with the best matching block of the second current tracking block, update the second current tracking block with the adjacent tracking block below the second current tracking block, and return to the step "Determine the adjacent matching block below the second reference block as the second intermediate matching block"; If the second judgment result is yes, the best matching blocks of all tracking blocks below the starting tracking block are determined.

4. The ultrasonic speckle tracking blood flow velocity measurement method according to claim 1, wherein Respectively determine the blood flow velocity corresponding to each tracking block according to the pixel coordinates of all tracking blocks and the pixel coordinates of all best matching blocks, specifically including: Respectively determine the displacement of each tracking block according to the pixel coordinates of all tracking blocks and the pixel coordinates of all best matching blocks; Respectively determine the blood flow velocity corresponding to each tracking block according to the displacements of all tracking blocks and the acquisition duration between the reference frame and the comparison frame.

5. The ultrasonic speckle tracking blood flow velocity measurement method according to claim 1, characterized in that Determine several tracking blocks along the position of the radial blood flow in the reference frame, and determine any tracking block in the middle area as the starting tracking block, specifically including: Determine several tracking blocks along the position of the radial blood flow in the reference frame; Number all tracking blocks in the order from the upper blood vessel wall to the lower blood vessel wall; Determine any tracking block with a serial number between 0.4W and 0.6W as the starting tracking block; where W is the total number of tracking blocks.

6. The ultrasonic speckle tracking blood flow velocity measurement method according to claim 1, characterized in that Also include: Determine the blood flow parabola according to the blood flow velocities corresponding to all tracking blocks.

7. An ultrasonic speckle tracking blood flow velocity measurement system, characterized in that, Include: An image acquisition module, configured to acquire an ultrasonic image sequence of a radial cross-section of a blood vessel, and determine a reference frame and a comparison frame from the ultrasonic image sequence; The comparison frame is the next frame image of the reference frame; Tracking block determination module, configured to determine a plurality of tracking blocks along the position where the radial blood flow is located in the reference frame, and determine any one of the tracking blocks in the middle area as the starting tracking block; Best matching block determination module for the starting tracking block, configured to determine the search area of the starting tracking block along the lateral direction inside the blood vessel in the comparison frame, and determine the best matching block of the starting tracking block within the search area of the starting tracking block; Best matching block determination module for the upper tracking blocks, configured to use the best matching block of the starting tracking block as a reference, and sequentially determine the best matching blocks of all the tracking blocks above the starting tracking block in the comparison frame in the order from bottom to top, specifically including: Determine the best matching block of the starting tracking block as the first reference block, and determine the adjacent tracking block above the starting tracking block as the first current tracking block; Determine the adjacent matching block above the first reference block as the first intermediate matching block; Taking the first intermediate matching block as the search starting point, adopt a search strategy with an adaptive search area, and determine the best matching block of the first current tracking block among several matching blocks on the left and right of the search starting point; the sum of the mean squared errors between the best matching block of the first current tracking block and the first current tracking block is simultaneously less than the sum of the mean squared errors between the matching block at a pixel point to the left of the best matching block of the first current tracking block and the first current tracking block and the sum of the mean squared errors between the matching block at a pixel point to the right of the best matching block of the first current tracking block and the first current tracking block; Judge whether the first current tracking block is the last tracking block above the starting tracking block, and obtain a first judgment result; If the first judgment result is negative, update the first reference block with the best matching block of the first current tracking block, update the first current tracking block with the adjacent tracking block above the first current tracking block, and return to the step of "determine the adjacent matching block above the first reference block as the first intermediate matching block"; If the first judgment result is positive, the best matching blocks of all the tracking blocks above the starting tracking block are determined; Best matching block determination module for the lower tracking blocks, configured to use the best matching block of the starting tracking block as a reference, and sequentially determine the best matching blocks of all the tracking blocks below the starting tracking block in the comparison frame in the order from top to bottom; Blood flow velocity determination module, configured to determine the blood flow velocity corresponding to each tracking block according to the pixel coordinates of all the tracking blocks and the pixel coordinates of all the best matching blocks respectively.

8. An electronic device, characterized in that, Comprising a memory and a processor, the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to execute the ultrasonic speckle tracking blood flow velocity measurement method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, It stores a computer program, and when the computer program is executed by the processor, it implements the ultrasonic speckle tracking blood flow velocity measurement method according to any one of claims 1 to 6.

Images