A blood flow velocity measurement method based on optical flow method regional tracking

Abstract

The application discloses a blood flow velocity measurement method based on a light flow method area tracking, relates to the blood flow velocity measurement technical field, and comprises the following steps: extracting multiple frames of blood vessel images from a blood vessel video, and determining a ROI area in each frame of the blood vessel images according to a predetermined ROI area division rule; performing movement estimation on adjacent two frames of the blood vessel images by using a DIS light flow method to obtain blood flow data of the adjacent two frames of the blood vessel images; performing area tracking on the blood flow data by using an affine transformation to obtain affine transformation parameters of each frame of the blood vessel images; inputting the affine transformation parameters into a preset movement amount prediction model, predicting movement amounts of ROI areas of multiple groups of adjacent frames of the blood vessel images, and performing velocity estimation according to the multiple groups of the movement amounts to output blood flow velocities; the application performs movement estimation on the adjacent two frames of the blood vessel images by using the DIS light flow method, improves the response capability of a system to dynamic changes of blood flow in a complex blood vessel structure, and thus ensures accurate measurement under different blood vessel conditions.

Description

A method for measuring vascular blood flow velocity based on optical flow region tracking Technical Field

[0001] This invention relates to the field of blood flow velocity measurement, and specifically to a method for measuring vascular blood flow velocity based on optical flow region tracking. Background Technology

[0002] Blood flow velocity in the microcirculation is crucial for maintaining tissue health. Slowed blood flow or thrombosis can lead to local tissue ischemia, hypoxia, and even necrosis, which can cause various diseases. Therefore, as an important part of blood circulation, blood flow velocity in the microcirculation is an important physiological parameter reflecting the state of the microcirculation. In order to accurately measure the blood flow velocity in the microcirculation, rapid and non-invasive modern blood flow measurement technology plays an increasingly important role in medical diagnosis.

[0003] Blood flow velocimetry technology has undergone years of development, evolving from early contact-based measurements to modern non-invasive measurement techniques based on blood flow imaging. Traditional contact methods often suffer from invasiveness, operational complexity, and low accuracy, while modern methods offer more reliable, rapid, and non-invasive options. Modern blood flow velocimetry technologies are mainly divided into three categories: frequency shift analysis methods based on the Doppler effect, analysis methods based on tracer particle tracking, and motion estimation methods based on image pixel features. The Doppler effect method is commonly used for blood flow velocimetry in large vessels, but its application in microcirculation and low-velocity blood flow is less advanced. Blood flow is easily affected by clutter interference; while tracer particle tracking is effective, tracer particles cannot be placed in the blood, thus limiting its practical application. In contrast, motion estimation methods based on image pixel features, which analyze the motion characteristics of red or white blood cells in the blood for velocity measurement, have become an important means of studying microcirculatory blood flow, especially widely used in non-invasive quantitative assessment of nailfold microvascular blood flow. For example, Chinese patent CN111956200B proposes a microcirculatory high-speed blood flow measurement system and method. Combining with existing technologies, it can be found that:

[0004] Currently, it is difficult to achieve high-precision blood flow velocity detection in complex vascular structures, especially when dealing with different vascular conditions and dynamic changes. The velocity measurement results are easily affected by factors such as non-target areas, noise, and poor image quality. In addition, the lack of an intelligent prediction mechanism based on historical data leads to insufficient accuracy of velocity estimation, and weak robustness and anti-interference ability. Therefore, there is a need for a technology that can accurately delineate ROI regions, stably capture blood flow regions, and optimize velocity prediction through historical movement data to improve the accuracy, continuity, and anti-interference ability of blood flow velocity measurement. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides a method for measuring vascular blood flow velocity based on optical flow region tracking.

[0006] A method for measuring vascular blood flow velocity based on optical flow region tracking, the method comprising:

[0007] Multiple frames of vascular images are extracted from vascular videos, and the ROI region in each frame of vascular image is determined according to the established ROI region division rules.

[0008] The motion estimation of two adjacent vascular images is performed using the DIS optical flow method to obtain blood flow data of the two adjacent vascular images.

[0009] Affine transformation is used to perform region tracking on blood flow data to obtain the affine transformation parameters for each frame of vascular image.

[0010] The affine transformation parameters are input into a preset motion prediction model to predict the motion of the ROI region in multiple adjacent frames of vascular images. Based on the multiple motion values, the velocity is estimated to output the blood flow velocity.

[0011] Furthermore, the logic for setting the ROI region division rules is as follows:

[0012] Select a target area with clear blood features in the blood vessel image as the ROI region. The size of the ROI region is determined according to the width of the blood vessel. The width of the ROI region shall not exceed the width of the narrowest part of the blood vessel. The ROI region is set in the middle of the blood vessel. Based on the determined position of the ROI region, multiple ROI regions are drawn.

[0013] Furthermore, the blood flow data includes the optical flow field of the vascular image, the horizontal velocity component of the optical flow, and the vertical velocity component of the optical flow.

[0014] Furthermore, motion estimation is performed on two adjacent frames of vascular images using the DIS optical flow method to obtain blood flow data, including:

[0015] The DIS optical flow algorithm is used to calculate the optical flow field for two adjacent frames of blood vessel images. The logical formula for the optical flow field is as follows: ;in, For optical flow field, For the horizontal velocity component of optical flow, This represents the vertical velocity component of the optical flow.

[0016] Furthermore, affine transformations are used to perform region tracking of blood flow data, including:

[0017] Optical flow within the Region of Interest (ROI) of the optical flow field in the blood flow data is extracted. An affine transformation is then applied to map the ROI region of the t-1 frame vascular image to the t frame vascular image to determine the affine transformation parameters, which include translation and scaling parameters. The range of the ROI region from the t-1 frame vascular image to the t frame vascular image is set to (…). , )arrive( , The fitting formula for horizontal motion is: ;in, Let be the horizontal translation component of the i-th ROI region, where i is an integer greater than 0. This represents the horizontal scaling component of the ROI region. Let be the translation parameter for the fit, and s be the scaling parameter for the fit. The error term for horizontal motion is given; the fitting formula for vertical motion is: ;in, This represents the vertical translation component of the ROI region. This represents the scaling component of the ROI region in the vertical direction. The error term for vertical motion is given; the translation parameter P and scaling parameter S are calculated using the least squares method through the fitting formulas for vertical and horizontal motion.

[0018] Furthermore, the generation logic of the mobility prediction model is as follows:

[0019] Q1. Obtain historical motion training data and divide the historical motion training data into a motion training set and a motion test set; the historical motion training data includes affine transformation parameters and their corresponding motion values;

[0020] Q2. Construct a regression network by using the affine transformation parameters in the motion training set as input data and the corresponding estimated motion in the motion training set as output data. Train the regression network to obtain the initial machine learning network.

[0021] Q3. Use the mobility test set to validate the initial machine learning network. The initial machine learning network whose output is less than or equal to the preset test error threshold is used as the pre-built mobility prediction model.

[0022] Furthermore, the logic for obtaining the movement amount is as follows:

[0023] Based on the updated ROI regions of two adjacent vascular images from historical data, let the two adjacent images be the (N-1)th frame vascular image and the Nth frame vascular image, obtain the ROI region of the (N-1)th frame vascular image and mark it as... The region is the ROI region of the N-1 frame blood vessel images that is updated in the Nth frame blood vessel image, and is marked as... The region, among which, The area can be fully covered Given a region where the content within the region remains the same, let the range of the ROI region from N-1 frames of vascular images to N frames of vascular images be ( , )arrive( , ), The update formula for the region is: ;in, Historical data for the horizontal scaling component. Historical data for the scaling component in the vertical direction. This is historical data for the horizontal translation component. Historical data for the vertical translation component; based on The centroid of the region and The centroid of the interior is calculated and obtained. The centroid of the region and The distance between the centroids within the space is used as the amount of movement.

[0024] Furthermore, velocity estimation is performed based on multiple sets of movement data to output blood flow velocity, including:

[0025] The frame rate of the blood vessel image is obtained, and the time interval between two adjacent frames is calculated. Based on time and movement, the first-order blood flow velocity is calculated. The calculation equations for the first-order blood flow velocity are as follows: The system of equations is transformed into a formula for first-order blood flow velocity: Where Tg is the time interval between two frames, and ft is the frame rate of the blood vessel image. R represents the first-order blood flow velocity, and R represents the predicted amount of movement.

[0026] Suppose that d groups of first-order blood flow velocities are generated based on the movement amount of d groups. Calculate the mean first-order blood flow velocity Vj of the d groups of first-order blood flow velocities, and set an error threshold K, where K is a constant greater than 0. Filter out groups z with values ​​greater than [a certain value]. and less than The primary blood flow velocity is obtained by averaging the primary blood flow velocities of the remaining dz groups to obtain the secondary blood flow velocity Vz. The secondary blood flow velocity is then output as the blood flow velocity. The direction of blood flow velocity is determined by the centroid of the ROI region in consecutive frames.

[0027] A vascular blood flow velocity measurement system based on optical flow region tracking, used to perform any one of the vascular blood flow velocity measurement methods based on optical flow region tracking, the system comprising:

[0028] ROI region segmentation module: used to extract multiple frames of blood vessel images from blood vessel videos, and determine the ROI region in each frame of blood vessel images according to the established ROI region segmentation rules;

[0029] Motion estimation module: Used to estimate the motion of two adjacent vascular images using the DIS optical flow method to obtain blood flow data from the two adjacent vascular images.

[0030] Affine tracking module: used to perform region tracking of blood flow data through affine transformation to obtain the affine transformation parameters of each frame of vascular image;

[0031] Velocity analysis module: It is used to input affine transformation parameters into a preset motion prediction model, predict the motion of the ROI region of multiple adjacent frames of vascular images, and estimate the velocity based on multiple motion values ​​to output the blood flow velocity.

[0032] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0033] This invention achieves high-precision detection of blood flow velocity in blood vessels by estimating motion between two adjacent frames of vascular images using the DIS optical flow method. This enhances the system's responsiveness to dynamic changes in blood flow within complex vascular structures, ensuring accurate measurements under various vascular conditions. Furthermore, this invention selects the central region of a blood vessel with prominent blood characteristics as the ROI region through pre-defined ROI region division rules, and limits the size of the ROI region based on the width of the narrowest point of the vessel. This ensures that the ROI region accurately represents the blood flow characteristics within the blood vessel, reducing interference from non-target regions on the calculation results. Affine transformation tracking of the ROI region achieves stable capture of the blood flow region, effectively improving the stability and continuity of blood flow velocity measurement, thereby enhancing its robustness and anti-interference capability. In addition, a motion prediction model is constructed using historical motion training data, and affine transformation parameters are optimized based on a regression network, enabling intelligent analysis of blood flow velocity. This further improves the accuracy of blood flow velocity prediction and reduces velocity estimation errors caused by noise or poor image quality.

[0034] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0035] Figure 1 is a flowchart of a blood flow velocity measurement method based on optical flow region tracking provided in Embodiment 1 of the present invention;

[0036] Figure 2 is a diagram illustrating ROI region division based on optical flow region tracking according to Embodiment 1 of the present invention;

[0037] Figure 3 is a schematic diagram of a region tracking result based on optical flow method provided in Embodiment 1 of the present invention;

[0038] Figure 4 is a block diagram of a vascular blood flow velocity measurement system based on optical flow region tracking provided in Embodiment 2 of the present invention. Detailed Description

[0039] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0040] Example 1

[0041] Please refer to Figure 1. This embodiment discloses a method for measuring vascular blood flow velocity based on optical flow region tracking. The method includes:

[0042] S110: Extract multiple frames of blood vessel images from the blood vessel video, and determine the ROI region in each frame of blood vessel image according to the predetermined ROI region division rules.

[0043] Specifically, referring to Figure 2, the logic for setting the ROI region division rules is as follows:

[0044] Select a target area with clear blood features in the blood vessel image as the ROI region. The size of the ROI region is determined according to the width of the blood vessel. The width of the ROI region shall not exceed the width of the narrowest part of the blood vessel. The ROI region is set in the middle of the blood vessel. Based on the determined position of the ROI region, multiple ROI regions are drawn.

[0045] It should be noted that blood flow characteristics include blood flow velocity, direction, flow rate, pulsation, laminar or turbulent flow state, etc. These data are mainly acquired through the following non-invasive acquisition methods: ultrasound Doppler uses the Doppler effect to measure blood flow velocity and direction; magnetic resonance angiography (MRA) and computed tomography angiography (CTA) combined with contrast agents provide three-dimensional vascular images to assess blood flow velocity and occlusion; near-infrared spectroscopy (NIRS) indirectly measures local blood flow through light absorption; and laser Doppler imaging (LDI) uses laser reflection frequency shift to measure microcirculatory blood flow.

[0046] S120 uses the DIS optical flow method to estimate the motion of two adjacent vascular images to obtain blood flow data of two adjacent vascular images.

[0047] Specifically, the blood flow data includes the optical flow field, horizontal velocity component, and vertical velocity component of the optical flow in the vascular image;

[0048] Specifically, the DIS optical flow method is used to perform motion calculations on two adjacent frames of vascular images to obtain blood flow data, including:

[0049] The DIS optical flow algorithm is used to calculate the optical flow field for two adjacent frames of blood vessel images. The logical formula for the optical flow field is as follows: ;in, For optical flow field, For the horizontal velocity component of optical flow, This represents the vertical velocity component of the optical flow.

[0050] It should be noted that the optical flow field is a 2D vector field, which assigns a predicted motion vector from the current frame to the next frame to each pixel in the blood vessel image. This vector describes the pixel displacement vector between the blood vessel image in frame t-1 and frame t, where the horizontal velocity component... (x,y) and vertical velocity components (x,y) is obtained by calculation using the optical flow equation.

[0051] S130 performs region tracking on blood flow data through affine transformation to obtain the affine transformation parameters of each frame of vascular image.

[0052] Specifically, affine transformation is used to perform region tracking of blood flow data, including:

[0053] Optical flow within the region of origin (ROI) of the optical flow field in the blood flow data is extracted. The ROI region of the t-1 frame blood vessel image is mapped to the t frame blood vessel image by fitting an affine transformation to determine the affine transformation parameters, which include translation parameters and scaling parameters.

[0054] It's important to understand that optical flow refers to the motion vector of each pixel in an image between two consecutive image frames. It describes the movement or change of a pixel over time.

[0055] Let the range of the ROI region from the t-1 frame vascular image to the t frame vascular image be ( , )arrive( , The fitting formula for horizontal motion is: ;in, Let be the horizontal translation component of the i-th ROI region, where i is an integer greater than 0. This represents the horizontal scaling component of the ROI region. Let be the translation parameter for the fit, and s be the scaling parameter for the fit. This is the error term for horizontal motion;

[0056] What needs to be understood is: error term Represents the horizontal velocity component of optical flow Displacement predicted by affine transformation model By minimizing the error term, the square of the difference between the two can be used to find the optimal translation parameter. and scaling parameters ;

[0057] The fitting formula for vertical motion is: ;in, This represents the vertical translation component of the ROI region. This represents the scaling component of the ROI region in the vertical direction. This is the error term for vertical motion;

[0058] What needs to be understood is: error term Represents the vertical velocity component of optical flow Displacement predicted by affine transformation model The square of the difference between them;

[0059] The translation parameter P and scaling parameter S are obtained by using the fitting formulas for vertical and horizontal motions and the least squares method.

[0060] It should be noted that when tracking a blood vessel region of interest (ROI), affine transformations such as rotation and scaling may occur within the ROI due to the deformation of the blood vessel fluid. By fitting the affine transformation, the new position of the blood vessel region can be accurately estimated. Since the calculation of blood flow velocity depends on the precise changes in position between consecutive frames, fitting the affine transformation is crucial for measuring blood flow velocity. The fitting of the affine transformation is to find the translation and scaling parameters that best suit the current ROI region motion, thereby effectively updating the position and size of the ROI region. By minimizing the error between the optical flow field and the affine transformation prediction, the affine transformation can provide a more accurate local motion estimate for the ROI region, further improving the accuracy of the tracking algorithm, especially when the ROI region undergoes rapid or complex motion.

[0061] S140, input the affine transformation parameters into the preset motion prediction model, predict the motion of the ROI region of multiple adjacent frame vascular images, and estimate the velocity based on the multiple motion values ​​to output the blood flow velocity.

[0062] Specifically, the generation logic of the mobility prediction model is as follows:

[0063] Q1. Obtain historical motion training data and divide the historical motion training data into a motion training set and a motion test set; the historical motion training data includes affine transformation parameters and their corresponding motion values;

[0064] Specifically, referring to Figure 3, the logic for obtaining the movement amount is as follows:

[0065] Based on the updated ROI regions of two adjacent vascular images from historical data, let the two adjacent images be the (N-1)th frame vascular image (upper vascular image in Figure 3) and the Nth frame vascular image (lower vascular image in Figure 3), obtain the ROI region of the (N-1)th frame vascular image and mark it as... The region is the ROI region of the N-1 frame blood vessel images that is updated in the Nth frame blood vessel image, and is marked as... The region, among which, The area can be fully covered Given a region where the content within the region remains the same, let the range of the ROI region from N-1 frames of vascular images to N frames of vascular images be ( , )arrive( , ), The update formula for the region is: ;in, Historical data for the horizontal scaling component. Historical data for the scaling component in the vertical direction. This is historical data for the horizontal translation component. Historical data for the vertical translation component;

[0066] It should be noted that the area within the box in Figure 3 is the ROI region. In the region update formula, for The new position of the coordinates within the region after translation and scaling. and They represent The width and height of the area after scaling;

[0067] according to The centroid of the region and The centroid of the interior is calculated and obtained. The centroid of the region and The distance between the centroids within the space is used as the amount of movement.

[0068] It should be noted that the meaning of 'i' in this step is the same as above, but in this step, 'i' refers to the i-th ROI region in the historical data, with its centroid at [value missing]. The regional center point within the region;

[0069] Q2. Construct a regression network by using the affine transformation parameters in the motion training set as input data and the corresponding estimated motion in the motion training set as output data. Train the regression network to obtain the initial machine learning network.

[0070] Q3. Use the mobility test set to validate the initial machine learning network. The initial machine learning network whose output is less than or equal to the preset test error threshold is used as the pre-built mobility prediction model.

[0071] Specifically, velocity is estimated based on multiple sets of movement data to output blood flow velocity, including:

[0072] The frame rate of the blood vessel image is obtained, and the time interval between two adjacent frames is calculated. Based on time and movement, the first-order blood flow velocity is calculated. The calculation equations for the first-order blood flow velocity are as follows: The system of equations is transformed into a formula for first-order blood flow velocity: Where Tg is the time interval between two frames, and ft is the frame rate of the blood vessel image. R represents the first-order blood flow velocity, and R represents the predicted amount of movement.

[0073] Suppose that d groups of first-order blood flow velocities are generated based on the movement amount of d groups. Calculate the mean first-order blood flow velocity Vj of the d groups of first-order blood flow velocities, and set an error threshold K, where K is a constant greater than 0. Filter out groups z with values ​​greater than [a certain value]. and less than The primary blood flow velocity is obtained by averaging the primary blood flow velocities of the remaining dz groups to obtain the secondary blood flow velocity Vz. The secondary blood flow velocity is then output as the blood flow velocity. The direction of blood flow velocity is determined by the centroid of the ROI region in consecutive frames.

[0074] It should be noted that the time interval Tx can be obtained through a timer.

[0075] Example 2

[0076] Please refer to Figure 3. Based on the unified inventive concept, this embodiment also discloses a vascular blood flow velocity measurement system based on optical flow region tracking. The system includes a ROI region division module, a motion estimation module, an affine tracking module, and a velocity analysis module.

[0077] Among them, the ROI region segmentation module S210 is used to extract multiple frames of blood vessel images from the blood vessel video and determine the ROI region in each frame of blood vessel image according to the established ROI region segmentation rules.

[0078] Specifically, the logic for setting the ROI region division rules is as follows:

[0079] Select a target area with clear blood features in the blood vessel image as the ROI region. The size of the ROI region is determined according to the width of the blood vessel. The width of the ROI region shall not exceed the width of the narrowest part of the blood vessel. The ROI region is set in the middle of the blood vessel. Based on the determined position of the ROI region, multiple ROI regions are drawn.

[0080] Motion estimation module S220: used to perform motion estimation on two adjacent vascular images using the DIS optical flow method to obtain blood flow data of the two adjacent vascular images;

[0081] Specifically, the blood flow data includes the optical flow field, horizontal velocity component, and vertical velocity component of the optical flow in the vascular image;

[0082] Specifically, the DIS optical flow method is used to perform motion calculations on two adjacent frames of vascular images to obtain blood flow data, including:

[0083] The DIS optical flow algorithm is used to calculate the optical flow field for two adjacent frames of blood vessel images. The logical formula for the optical flow field is as follows: ;in, For optical flow field, For the horizontal velocity component of optical flow, This represents the vertical velocity component of the optical flow.

[0084] Affine tracking module S230: used to perform region tracking of blood flow data through affine transformation to obtain the affine transformation parameters of each frame of vascular image;

[0085] Specifically, affine transformation is used to perform region tracking of blood flow data, including:

[0086] Optical flow within the region of origin (ROI) of the optical flow field in the blood flow data is extracted. The ROI region of the t-1 frame blood vessel image is mapped to the t frame blood vessel image by fitting an affine transformation to determine the affine transformation parameters, which include translation parameters and scaling parameters.

[0087] Let the range of the ROI region from the t-1 frame vascular image to the t frame vascular image be ( , )arrive( , The fitting formula for horizontal motion is: ;in, Let be the horizontal translation component of the i-th ROI region, where i is an integer greater than 0. This represents the horizontal scaling component of the ROI region. Let be the translation parameter for the fit, and s be the scaling parameter for the fit. This is the error term for horizontal motion;

[0088] The fitting formula for vertical motion is: ;in, This represents the vertical translation component of the ROI region. This represents the scaling component of the ROI region in the vertical direction. This is the error term for vertical motion;

[0089] The translation parameter P and scaling parameter S are obtained by using the fitting formulas for vertical and horizontal motions and the least squares method.

[0090] Velocity analysis module S240: It is used to input affine transformation parameters into a preset motion prediction model, predict the motion of the ROI region of multiple adjacent frames of blood vessel images, and perform velocity estimation based on multiple motion values ​​to output blood flow velocity.

[0091] Specifically, the generation logic of the mobility prediction model is as follows:

[0092] Q1. Obtain historical motion training data and divide the historical motion training data into a motion training set and a motion test set; the historical motion training data includes affine transformation parameters and their corresponding motion values;

[0093] Specifically, the logic for obtaining the movement amount is as follows:

[0094] Based on the updated ROI regions of two adjacent vascular images from historical data, let the two adjacent images be the (N-1)th frame vascular image and the Nth frame vascular image, obtain the ROI region of the (N-1)th frame vascular image and mark it as... The region is the ROI region of the N-1 frame blood vessel images that is updated in the Nth frame blood vessel image, and is marked as... The region, among which, The area can be fully covered Given a region where the content within the region remains the same, let the range of the ROI region from N-1 frames of vascular images to N frames of vascular images be ( , )arrive( , ), The update formula for the region is: ;in, Historical data for the horizontal scaling component. Historical data for the scaling component in the vertical direction. This is historical data for the horizontal translation component. Historical data for the vertical translation component;

[0095] according to The centroid of the region and The centroid of the interior is calculated and obtained. The centroid of the region and The distance between the centroids within the space is used as the amount of movement.

[0096] Q2. Construct a regression network by using the affine transformation parameters in the motion training set as input data for the regression network and the estimated motion in the motion training set as output data for the regression network. Train the regression network to obtain the initial machine learning network.

[0097] Q3. Use the mobility test set to validate the initial machine learning network. The initial machine learning network whose output is less than or equal to the preset test error threshold is used as the pre-built mobility prediction model.

[0098] Specifically, velocity is estimated based on multiple sets of movement data to output blood flow velocity, including:

[0099] The frame rate of the blood vessel image is obtained, and the time interval between two adjacent frames is calculated. Based on time and movement, the first-order blood flow velocity is calculated. The calculation equations for the first-order blood flow velocity are as follows: The system of equations is transformed into a formula for first-order blood flow velocity: Where Tg is the time interval between two frames, and ft is the frame rate of the blood vessel image. R represents the first-order blood flow velocity, and R represents the predicted amount of movement.

[0100] Suppose that d groups of first-order blood flow velocities are generated based on the movement amount of d groups. Calculate the mean first-order blood flow velocity Vj of the d groups of first-order blood flow velocities, and set an error threshold K, where K is a constant greater than 0. Filter out groups z with values ​​greater than [a certain value]. and less than The primary blood flow velocity is obtained by averaging the primary blood flow velocities of the remaining dz groups to obtain the secondary blood flow velocity Vz. The secondary blood flow velocity is then output as the blood flow velocity. The direction of blood flow velocity is determined by the centroid of the ROI region in consecutive frames.

[0101] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via a wired or wireless network. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0102] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed in this invention can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0103] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0104] In the several embodiments provided by this invention, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of units is only for one method of vascular blood flow velocity measurement based on optical flow region tracking. In actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces, or indirect coupling or communication connection between devices or units, and may be electrical, mechanical, or other forms.

[0105] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0106] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0107] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

[0108] In conclusion, the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for measuring vascular blood flow velocity based on optical flow region tracking, characterized in that, The method includes: extracting multiple frames of vascular images from a vascular video; determining the Region of Interest (ROI) in each frame of the vascular image according to a predetermined ROI region division rule; the logic for setting the ROI region division rule is as follows: selecting a target area with clear blood features in the vascular image as the ROI region; the size of the ROI region is determined according to the width of the vascular image; the width of the ROI region does not exceed the width of the narrowest part of the vascular image; the ROI region is set in the middle of the vascular image; and drawing multiple ROI regions based on the determined ROI region positions; performing motion estimation on two adjacent frames of vascular images using the DIS optical flow method to obtain blood flow data of the two adjacent frames of vascular images; the blood flow data includes the optical flow field, horizontal velocity component, and vertical velocity component of the optical flow in the vascular image; and calculating the optical flow field using the DIS optical flow algorithm on two adjacent frames of vascular images, the logical formula for the optical flow field being: ;in, For optical flow field, For the horizontal velocity component of optical flow, The optical flow is represented by the vertical velocity component. Affine transformation is used to perform region tracking on the blood flow data to obtain the affine transformation parameters for each frame of the vascular image. This includes: extracting the optical flow within the region of origin (ROI) of the optical flow field in the blood flow data; mapping the ROI region of frame t-1 of the vascular image to frame t of the vascular image by fitting an affine transformation to determine the affine transformation parameters, which include translation and scaling parameters; the range of the ROI region from frame t-1 to frame t of the vascular image is defined as (…). ， )arrive( ， The fitting formula for horizontal motion is: ;in, Let be the horizontal translation component of the i-th ROI region, where i is an integer greater than 0. This represents the horizontal scaling component of the ROI region. Let be the translation parameter for the fit, and s be the scaling parameter for the fit. The error term for horizontal motion is given; the fitting formula for vertical motion is: ;in, This represents the vertical translation component of the ROI region. This represents the scaling component of the ROI region in the vertical direction. The error term for vertical motion is calculated using the fitting formulas for vertical and horizontal motion, and the translation parameter P and scaling parameter S are obtained by least squares method. The affine transformation parameters are input into the preset motion prediction model to predict the motion of the ROI region of multiple adjacent frames of blood vessel images, and the velocity is estimated based on the multiple motion values ​​to output the blood flow velocity.

2. The method for measuring vascular blood flow velocity based on optical flow region tracking according to claim 1, characterized in that, The generation logic of the mobility prediction model is as follows: Q1, obtain historical mobility training data and divide the historical mobility training data into a mobility training set and a mobility test set; the historical mobility training data includes affine transformation parameters and their corresponding mobility; Q2, construct a regression network, use the affine transformation parameters in the mobility training set as the input data of the regression network, and use the estimated mobility corresponding to the mobility training set as the output data of the regression network, train the regression network to obtain the initial machine learning network; Q3. Use the mobility test set to validate the initial machine learning network. The initial machine learning network whose output is less than or equal to the preset test error threshold is used as the pre-built mobility prediction model.

3. The method for measuring vascular blood flow velocity based on optical flow region tracking according to claim 2, characterized in that, The logic for obtaining the movement amount is as follows: Based on the updated ROI region data of two adjacent frames of vascular images in history, assuming the two adjacent frames are the (N-1)th frame vascular image and the Nth frame vascular image, the ROI region of the (N-1)th frame vascular image is obtained and marked as... The region is the ROI region of the N-1 frame blood vessel images that is updated in the Nth frame blood vessel image, and is marked as... The region, among which, The area can be fully covered Given a region where the content within the region remains the same, let the range of the ROI region from N-1 frames of vascular images to N frames of vascular images be ( ， )arrive( ， ）， The update formula for the region is: ;in, Historical data for the horizontal scaling component. Historical data for the scaling component in the vertical direction. This is historical data for the horizontal translation component. Historical data for the vertical translation component; based on The centroid of the region and The centroid of the interior is calculated and obtained. The centroid of the region and The distance between the centroids within the space is used as the amount of movement.

4. The method for measuring vascular blood flow velocity based on optical flow region tracking according to claim 3, characterized in that, Velocity estimation is performed based on multiple sets of movement data to output blood flow velocity. This includes: acquiring the frame rate of vascular images, calculating the time interval between two adjacent frames, and calculating the first-order blood flow velocity based on time and movement data. The calculation equations for the first-order blood flow velocity are as follows: The system of equations is transformed into a formula for first-order blood flow velocity: Where Tg is the time interval between two frames, and ft is the frame rate of the blood vessel image. Let R be the first-order blood flow velocity, and R be the predicted movement amount. Assume that d groups of first-order blood flow velocities are generated based on the movement amounts of d groups. Calculate the mean first-order blood flow velocity Vj of the d groups of first-order blood flow velocities. Set an error threshold K, where K is a constant greater than 0. Filter out groups z with values ​​greater than [value missing]. and less than The primary blood flow velocity is obtained by averaging the primary blood flow velocities of the remaining dz groups to obtain the secondary blood flow velocity Vz. The secondary blood flow velocity is then output as the blood flow velocity. The direction of blood flow velocity is determined by the centroid of the ROI region in consecutive frames.

5. A vascular blood flow velocity measurement system based on optical flow region tracking, used to execute the vascular blood flow velocity measurement method based on optical flow region tracking as described in any one of claims 1-4, characterized in that, The system includes: a Region of Interest (ROI) segmentation module for extracting multiple frames of vascular images from a vascular video and determining the ROI region in each frame of the vascular image according to a predetermined ROI segmentation rule; a motion estimation module for performing motion estimation on two adjacent frames of vascular images using the DIS optical flow method to obtain blood flow data of the two adjacent frames of vascular images; an affine tracking module for performing region tracking on blood flow data through affine transformation to obtain affine transformation parameters for each frame of vascular image; and a velocity analysis module for inputting the affine transformation parameters into a preset motion prediction model to predict the motion of the ROI region in multiple sets of adjacent frames of vascular images, and performing velocity estimation based on the multiple sets of motion to output the blood flow velocity.

Images