Method and computer device for determining blood vessel flow velocity

Abstract

The embodiment of the present application discloses a method and computer device for determining blood vessel flow velocity. The method includes: acquiring angiography sequence images, finding a coronary angiography image that meets the best development conditions in the angiography sequence images as the best coronary angiography image, selecting the coronary angiography image arranged before the timestamp of the best coronary angiography image in the angiography sequence images in sequence as the target coronary angiography image, the target coronary angiography image and the best coronary angiography image forming an image segmentation set, generating a blood vessel length change graph based on the image segmentation set, determining the conduction start time and the conduction end time based on the blood vessel length change graph, and determining the blood vessel flow velocity based on the main branch vessel length, the conduction start time and the conduction end time. The embodiment of the present application does not need to analyze all coronary angiography images, and selects the best coronary angiography image and the target coronary angiography image arranged before the best coronary angiography image for analysis, which can reduce the number of blood vessel segmentations.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of medical image processing technology, and specifically to a method for determining blood vessel flow velocity and a computer device. Background Art

[0002] Vascular flow velocity is of great significance for disease diagnosis. When determining vascular flow velocity, the relevant technology needs to obtain all coronary angiography images during the vascular development process, and segment and analyze all coronary angiography images to effectively determine the vascular flow velocity. Therefore, this method takes a lot of time and the detection efficiency of vascular flow velocity is low. Summary of the invention

[0003] One purpose of the embodiments of the present application is to provide a method for determining blood vessel flow velocity and a computer device to solve the technical problem of low efficiency in detecting blood vessel flow velocity in the related art.

[0004] In a first aspect, an embodiment of the present application provides a method for determining a blood vessel flow velocity, comprising:

[0005] Acquiring angiography sequence images, wherein the angiography sequence images include multiple frames of coronary angiography images;

[0006] Searching out a coronary angiography image that meets the best development condition in the angiography sequence images as the best coronary angiography image;

[0007] Selecting in sequence from the angiography sequence images a coronary angiography image arranged before the timestamp of the best coronary angiography image as a target coronary angiography image, wherein the target coronary angiography image and the best coronary angiography image form an image segmentation set;

[0008] generating a blood vessel length change map based on the image segmentation set, wherein the blood vessel length change map is used to represent the change of the main branch blood vessel length with conduction time;

[0009] Determine the conduction start time and conduction end time based on the blood vessel length change graph;

[0010] The vascular flow velocity is determined based on the main branch vessel length, conduction start time and conduction end time.

[0011] Optionally, generating a blood vessel length change map based on the image segmentation set includes:

[0012] Performing blood vessel segmentation processing on the optimal coronary angiography image to obtain an optimal blood vessel segmentation map;

[0013] Performing blood vessel segmentation processing on the target coronary angiography image to obtain a target blood vessel segmentation map, wherein the target blood vessel segmentation map and the optimal blood vessel segmentation map constitute a blood vessel change sequence;

[0014] A blood vessel length change map is generated based on the blood vessel change sequence.

[0015] Optionally, performing blood vessel segmentation processing on the target coronary angiography image to obtain a target blood vessel segmentation map includes:

[0016] Performing blood vessel segmentation processing on the target coronary angiography image to obtain a candidate blood vessel segmentation map;

[0017] The candidate blood vessel segmentation map is subjected to noise suppression processing based on the optimal blood vessel segmentation map to obtain a target blood vessel segmentation map.

[0018] Optionally, performing noise suppression processing on the candidate blood vessel segmentation map based on the optimal blood vessel segmentation map to obtain a target blood vessel segmentation map includes:

[0019] Determine a reference blood vessel segmentation map, where the reference blood vessel segmentation map is a blood vessel segmentation map adjacent to the candidate blood vessel segmentation map and arranged after the candidate blood vessel segmentation map;

[0020] determining a blood vessel intersection region between the candidate blood vessel segmentation map and the reference blood vessel segmentation map;

[0021] The blood vessel intersection region is mapped to the optimal blood vessel segmentation map to obtain a target blood vessel segmentation map.

[0022] Optionally, determining a reference blood vessel segmentation map includes:

[0023] In response to the target coronary angiography image being a coronary angiography image adjacent to the best coronary angiography image, determining the best vessel segmentation map as a reference vessel segmentation map; or,

[0024] In response to the target coronary angiography image not being a coronary angiography image adjacent to the optimal coronary angiography image, a candidate vessel segmentation map of a coronary angiography image adjacent to the target coronary angiography image and arranged after the target coronary angiography image is determined as a reference vessel segmentation map.

[0025] Optionally, the determination method further includes:

[0026] In response to the sum of pixels of the target blood vessel segmentation image being less than or equal to a preset pixel threshold, or all coronary angiography images of the image segmentation set being in a traversed state, the blood vessel change sequence is output.

[0027] Optionally, searching for a coronary angiography image that satisfies an optimal development condition as an optimal coronary angiography image in the angiography sequence images includes:

[0028] Dividing all the coronary angiography images of the angiography sequence images into a plurality of image processing sets, each of the image processing sets including a plurality of mutually adjacent frames of coronary angiography images;

[0029] Using the image processing set as an input of a pre-trained blood vessel image processing model to obtain a maximum probability value corresponding to the image processing set;

[0030] Finding the maximum probability value of the maximum value among the maximum probability values ​​of all the image processing sets as the target probability value;

[0031] It is determined that the coronary angiography image corresponding to the target probability value meets the optimal development condition, and the coronary angiography image corresponding to the target probability value is the optimal coronary angiography image.

[0032] Optionally, the blood vessel length change graph is configured with a blood vessel length change curve, and the blood vessel length change curve is used to represent the change of the main branch blood vessel length with the conduction time, and determining the conduction start time and the conduction end time based on the blood vessel length change graph includes:

[0033] According to a preset segmentation length, the blood vessel length variation curve is segmented into a plurality of area segments;

[0034] Counting the length of blood vessels in each of the regional segments;

[0035] Generating a length histogram according to the blood vessel length of each of the region segments, wherein the length histogram is used to represent the distribution of the blood vessel lengths of each of the region segments;

[0036] The conduction start time and the conduction end time are determined based on the length histogram.

[0037] Optionally, the region segment is constrained by a left endpoint distance and a right endpoint distance, and determining the conduction start time and the conduction end time based on the length histogram includes:

[0038] Searching out a region segment satisfying a start frame condition from among a plurality of region segments of the length histogram as a start frame segment;

[0039] Searching out a region segment satisfying an end frame condition from among a plurality of region segments of the length histogram as an end frame segment;

[0040] Based on the segment range corresponding to the start frame segment, find out the conduction start time in the blood vessel length change graph;

[0041] Based on the segment range corresponding to the end frame segment, the conduction end time is found in the blood vessel length change graph.

[0042] In a second aspect, an embodiment of the present application provides a computer device, comprising a memory and a processor, wherein the memory is connected to the processor, and the processor is used to execute one or more computer programs stored in the memory. When the processor executes the one or more computer programs, the computer device implements the above-mentioned method for determining the vascular flow rate.

[0043] The embodiments of the present application can achieve the following technical effects: the embodiments of the present application do not need to analyze all coronary angiography images, but select the best coronary angiography image and the target coronary angiography image arranged before the best coronary angiography image for analysis. This can reduce the number of blood vessel segmentations, which is beneficial to improving the detection efficiency of blood vessel flow velocity, and can accurately determine the time when the contrast agent development effect is best. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the description of the embodiments of the present application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0045] Figure 1 A schematic diagram of the system architecture of a system for determining blood vessel flow velocity provided in an embodiment of the present application;

[0046] Figure 2 A schematic flow chart of a method for determining blood vessel flow velocity provided in an embodiment of the present application;

[0047] Figure 3 A schematic diagram of dividing all coronary angiography images of angiography sequence images into multiple image processing sets provided in an embodiment of the present application;

[0048] Figure 4 A schematic diagram of the architecture of a vascular image processing model provided in an embodiment of the present application;

[0049] Figure 5 A schematic diagram of a scenario of performing noise suppression processing on a coronary angiography image provided in an embodiment of the present application;

[0050] Figure 6 A schematic diagram of a blood vessel length change diagram provided in an embodiment of the present application;

[0051] Figure 7 A schematic diagram of a length histogram provided in an embodiment of the present application;

[0052] Figure 8 A schematic diagram of the structure of a device for determining blood vessel flow velocity provided in an embodiment of the present application;

[0053] Fig. 9 A schematic diagram of the structure of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0054] In order to make the purpose, technical solutions and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present application.

[0055] It should be noted that, if there is no conflict, the various features in the embodiments of the present application can be combined with each other, all within the scope of protection of the present application. In addition, although the functional module division is performed in the device schematic diagram and the logical order is shown in the flow chart, in some cases, the steps shown or described can be performed in a sequence different from the module division in the device or the flow chart. Furthermore, the words "first", "second", "third", etc. used in this application do not limit the data and execution order, but only distinguish the same items or similar items with basically the same functions and effects.

[0056] In addition to the technical defects of the related art pointed out in the background art, the inventors also found the following technical problems in other related technologies: When the related technology detects the blood vessel flow velocity, it is necessary to segment the main branch blood vessels from the coronary angiography image and determine the blood vessel flow velocity based on the center line of the main branch blood vessels, but the blood vessel flow velocity is easily affected by the image segmentation accuracy and the position of the center line. The image segmentation accuracy is easily affected by the accuracy of the blood vessel segmentation model and other aspects. The extraction of the center line is also affected by the image segmentation effect. When the segmented blood vessels are offset, not smooth enough or interrupted, the segmented blood vessels deviate from the blood vessels in the real position, thereby affecting the accuracy of determining the blood vessel length and blood vessel flow velocity.

[0057] The inventors also found that when the related technology determines the vascular flow rate, it is necessary to determine the TIMI (thrombolysis in myocardial infarction) frame number. The related technology uses a deep learning model to predict all coronary angiography images in order to obtain the TIMI frame number. This approach requires training to generate a deep learning model in advance, and it also takes a lot of time and computing power to obtain the TIMI frame number. The TIMI myocardial perfusion frame counting method refers to: calculating the time for the contrast agent to enter and exit the myocardium through the angiography frame number to quantitatively evaluate the myocardial perfusion.

[0058] Based on this, the embodiment of the present application provides a system for determining blood vessel flow velocity. The system 100 for determining blood vessel flow velocity includes a medical image acquisition device 11 and a computer device 12. The medical image acquisition device 11 is connected in communication with the computer device 12, wherein the communication connection includes a wired communication connection or a wireless communication connection, wherein the wired communication connection includes various communication connections that transmit information using tangible media such as metal wires and optical fibers. The wireless communication connection includes 6G communication, 5G communication, 4G communication, 3G communication, 2G communication or WIFI, etc.

[0059] The medical image acquisition device 11 is used to acquire coronary angiography images. The medical image acquisition device 11 may be a computed tomography (CT) device, a magnetic resonance imaging (MRI) device, or the like.

[0060] The computer device 12 processes the coronary angiography image according to the method for determining the blood vessel flow velocity provided in the following embodiments. It can be understood that the computer device 12 includes a desktop computer, a server or a mobile terminal.

[0061] First, the computer device 12 does not need to analyze all the coronary angiography images, but selects the best coronary angiography image and the target coronary angiography image arranged before the best coronary angiography image for analysis. This can reduce the number of blood vessel segmentations, help improve the detection efficiency of blood vessel flow velocity, and accurately determine the time when the contrast agent development effect is best.

[0062] Secondly, the computer device 12 uses the best coronary angiography image to suppress the noise of the target coronary angiography image, which is conducive to obtaining a more accurate and reliable target blood vessel segmentation map, and further can obtain an accurate and reliable main branch blood vessel length.

[0063] Finally, the computer device 12 does not need to use a deep learning model to predict the TIMI frame number (i.e., the difference between the conduction start time and the conduction end time), but only identifies the conduction start time and the conduction end time based on the change in the length of the main branch blood vessel, so that the TIMI frame number can be determined simply and quickly.

[0064] The following embodiments of the present application provide a method for determining blood vessel flow velocity. Figure 2 The method for determining the blood vessel flow velocity includes steps S21 to S26, which are specifically as follows:

[0065] The embodiment of the present application executes step S21 to obtain angiography sequence images, which include multiple frames of coronary angiography images. The coronary angiography image is an image obtained by scanning the target coronary artery after contrast agent is injected, and the coronary angiography image complies with the DICOM (Digital Imaging and Communications in Medicine) standard. The DICOM standard is a standard designed specifically for medical imaging, which is used to store, transmit and display medical images and is widely used in CT, MRI and hospital PACS (Picture Archiving and Communication Systems) systems.

[0066] There are many ways to obtain contrast imaging sequence images.

[0067] In some embodiments, the present application embodiment requests the medical image acquisition device to return angiography sequence images.

[0068] In some embodiments, the contrast sequence images are pre-stored locally on the computer device, and the embodiment of the present application controls the computer device to retrieve the contrast sequence images locally.

[0069] In some embodiments, the angiography sequence images are pre-stored on a third-party device, the third-party device is communicatively connected to the computer device, and the computer device requests the third-party device to return the angiography sequence images. The third-party device may be a cloud server or an edge device, etc.

[0070] In some embodiments, the contrast sequence images are pre-stored on a storage device, and when the storage device is inserted into a computer device, the computer device retrieves the contrast sequence images from the storage device.

[0071] In the embodiment of the present application, step S22 is performed to find the coronary angiography image that meets the best development condition in the angiography sequence images as the best coronary angiography image. The best coronary angiography image is the coronary angiography image when the contrast agent of the heart is fully developed at the end of diastole.

[0072] Finding a coronary angiography image that meets the best development condition in the angiography sequence images as the best coronary angiography image includes the following steps:

[0073] S221, dividing all the coronary angiography images in the angiography sequence into a plurality of image processing sets, each of which includes a plurality of mutually adjacent frames of coronary angiography images.

[0074] S222, using the image processing set as an input of the pre-trained blood vessel image processing model to obtain a maximum probability value corresponding to the image processing set.

[0075] S223, searching for a maximum probability value with a maximum value among the maximum probability values ​​of all image processing sets as a target probability value.

[0076] S224, determining that the coronary angiography image corresponding to the target probability value meets the optimal development condition, and the coronary angiography image corresponding to the target probability value is the optimal coronary angiography image.

[0077] In S221, the embodiment of the present application uses a specified number of adjacent coronary angiography images as an image processing set, and there are a target number of repeated coronary angiography images in two adjacent image processing sets. The specified number and the target number can be customized by the designer based on engineering experience. For example, the specified number is 3 and the target number is 2, that is, 3 adjacent frames of coronary angiography images are used as an image processing set, and there are 2 repeated coronary angiography images in two adjacent image processing sets.

[0078] See also Figure 3 The angiography sequence images include coronary angiography images F1 to coronary angiography images Fm arranged in chronological order, coronary angiography images F1 to coronary angiography images F3 constitute a first image processing set Q1, coronary angiography images F2 to coronary angiography images F4 constitute a second image processing set Q2, and there are repeated coronary angiography images F2 and coronary angiography images F3 in the first image processing set Q1 and the second image processing set Q2, and so on.

[0079] In S222, the image processing set is used as the input of the pre-trained vascular image processing model to obtain the maximum probability value corresponding to the image processing set, including the following steps: the image processing set is used as the input of the pre-trained vascular image processing model, so that the vascular image processing model outputs a probability value corresponding to each frame of the coronary angiography image in the image processing set, and the maximum probability value is found among multiple probability values.

[0080] For example, for the first image processing set Q1, coronary angiography image F1, coronary angiography image F2 and coronary angiography image F3 are input into the vascular image processing model as three-channel images. The vascular image processing model outputs a probability value p1 corresponding to the coronary angiography image F1, a probability value p2 corresponding to the coronary angiography image F2 and a probability value p3 corresponding to the coronary angiography image F3. The maximum probability value pmax0 is found among the probability values ​​p1, p2 and p3, and the maximum probability value pmax0 is used as the maximum probability value of the first image processing set Q1.

[0081] The embodiment of the present application generates a vascular image processing model based on deep learning algorithm training, wherein the deep learning algorithm includes ResNet algorithm or Yolo series algorithm, etc. Exemplarily, the embodiment of the present application generates a vascular image processing model based on ResNet algorithm training. In the training stage, the embodiment of the present application performs data collection to obtain multiple coronary angiography samples, and selects three consecutive coronary angiography samples for labeling each time, wherein the coronary angiography sample with the best development effect is set as a positive sample, and the remaining coronary angiography samples are negative samples, and the vascular image processing model is trained based on the positive samples and the negative samples.

[0082] See also Figure 4 The vascular image processing model includes an initial convolution layer 41, a maximum pooling layer 42, a ResNet network 43, an average pooling layer 44 and a fully connected layer 45. The embodiment of the present application compresses the size of each frame of the coronary angiography image of the image processing set to a specified size. Exemplarily, the specified size is 128*128.

[0083] The initial convolution layer 41 compresses and deepens the 3-channel coronary angiography image of size 128*128 to obtain a 64*64 vascular feature matrix of 64 channels. The maximum pooling layer 42 compresses the 64*64 vascular feature matrix of 64 channels to obtain a 32*32 vascular feature matrix of 64 channels.

[0084] The ResNet network 43 includes a plurality of ResNet modules, each of which includes a 2D convolution layer, a batch normalization (Batch Normalization, BatchNorm) layer and a ReLU activation layer. The 2D convolution layer is used to extract the vascular feature matrix. The batch normalization layer is used to normalize the output of the 2D convolution layer to obtain a normalized vascular feature matrix to improve the training speed and stability of the vascular image processing model. The ReLU activation layer performs a nonlinear transformation on the normalized vascular feature matrix to increase the nonlinear expression capability of the vascular image processing model.

[0085] like Figure 4As shown, the ResNet network 43 includes 16 ResNet modules, and the ResNet module set 431 does not change the dimension of the input vascular feature matrix, wherein the ResNet module set 431 includes the first ResNet module, the second ResNet module, and the third ResNet module. The ResNet module set 432 is used to compress and deepen the input of the ResNet module set 431, wherein the ResNet module set 432 includes the fourth ResNet module to the seventh ResNet module. The ResNet module set 433 is used to compress and deepen the input of the ResNet module set 432, wherein the ResNet module set 433 includes the eighth ResNet module to the thirteenth ResNet module. The ResNet module set 434 is used to compress and deepen the input of the ResNet module set 433, and finally output 512 4*4 vascular feature matrices, wherein the ResNet module set 434 includes the 14th ResNet module to the 16th ResNet module.

[0086] The average pooling layer 44 is used to fuse the 4*4 blood vessel feature matrix into a single-value feature for output.

[0087] The fully connected layer 45 is used to fuse the 512-dimensional single-value features into a single-value output, that is, to obtain a probability value.

[0088] The embodiment of the present application uses multiple adjacent frames of coronary angiography images as the input of the vascular image processing model. Since the multiple frames of coronary angiography images are relatively continuous in time, the vascular image processing model can better understand the dynamic changes in the angiography sequence images, which is beneficial to improving the accuracy of feature extraction.

[0089] In S223, the embodiment of the present application searches for the maximum probability value of the maximum value in {pmax0, pmax1, pmax2, ..., pmaxi, ... pmaxn} as the target probability value, where pmaxi is the maximum probability value of the i-th image processing set.

[0090] In S224, the target probability value represents that the coronary angiography image corresponding to the target probability value is the best coronary angiography image with the highest confidence among all coronary angiography images. Therefore, the embodiment of the present application regards the coronary angiography image corresponding to the target probability value as the best coronary angiography image.

[0091] The embodiment of the present application executes step S23 to sequentially select the coronary angiography images arranged before the timestamp of the best coronary angiography image in the angiography sequence images as the target coronary angiography images, and the target coronary angiography images and the best coronary angiography image constitute an image segmentation set.

[0092] For example, the best coronary angiography image is coronary angiography image F15, and the timestamps of coronary angiography images F1 to F14 are all before coronary angiography image F15, so coronary angiography images F1 to F14 are all target coronary angiography images. Coronary angiography images F1 to F15 constitute the image segmentation set W. In the embodiment of the present application, it is not necessary to analyze all coronary angiography images, but to select the best coronary angiography image and the target coronary angiography image arranged before the best coronary angiography image for analysis, which can reduce the number of blood vessel segmentation times and some length recording operations, and can accurately determine the time when the contrast agent development effect is best.

[0093] The embodiment of the present application executes step S24 to generate a blood vessel length change map based on the image segmentation set, where the blood vessel length change map is used to represent changes in the length of the main and branch blood vessels with conduction time.

[0094] The main branch vessel length is the length of the main branch vessel determined based on the image segmentation set, and the conduction time is the time for the contrast agent to be conducted, wherein the unit of the conduction time is the number of frames.

[0095] Generating a blood vessel length change map based on the image segmentation set includes the following steps:

[0096] S241, performing blood vessel segmentation processing on the best coronary angiography image to obtain an optimal blood vessel segmentation map.

[0097] S242, performing blood vessel segmentation processing on the target coronary angiography image to obtain a target blood vessel segmentation map, and the target blood vessel segmentation map and the optimal blood vessel segmentation map constitute a blood vessel change sequence.

[0098] S243, generating a blood vessel length change graph based on the blood vessel change sequence.

[0099] In S241, the best vessel segmentation map is an image of the main branch vessels segmented from the best coronary angiography image. In the embodiment of the present application, the target vessel type is determined, a segmentation model corresponding to the target vessel type is selected from a preset segmentation model library as a target segmentation model, and vessel segmentation processing is performed on the best coronary angiography image based on the target segmentation model to obtain the best vessel segmentation map.

[0100] Determining the target blood vessel type includes the following steps: presenting a blood vessel type selection page, the blood vessel type selection page including multiple blood vessel types, responding to a selection operation received on the blood vessel type selection page, and determining that the blood vessel type corresponding to the selection operation is the target blood vessel type.

[0101] The segmentation model library includes multiple segmentation models, and the segmentation models correspond to the blood vessel types. The blood vessel types include the right coronary artery (RCA), the left anterior descending coronary artery (LAD), and the left circumflex coronary artery (LCX). Correspondingly, the segmentation models include the right coronary artery model, the left anterior descending coronary artery model, and the left circumflex coronary artery model.

[0102] The segmentation model is a model trained by a deep learning algorithm, and is used to segment blood vessels from a coronary angiography image to obtain a blood vessel segmentation map. Exemplarily, the segmentation model is a U-Net segmentation model.

[0103] In S242, the target blood vessel segmentation map is an image of main branch blood vessels segmented from the target coronary angiography image. The embodiment of the present application performs blood vessel segmentation processing on the target coronary angiography image based on the target segmentation model to obtain the target blood vessel segmentation map.

[0104] Usually, the beating of the heart can easily cause the main branch blood vessels to shift in the coronary angiography images at different development times, or the angiography is not clear, or is affected by the segmentation accuracy of the segmentation model. The above situations can easily lead to problems such as virtual connections and blood vessel interruptions in the blood vessels segmented by the segmentation model, or it is easy to segment the blood vessels incorrectly. The above situations are all noise, and the embodiments of the present application require noise suppression processing of the coronary angiography images.

[0105] Performing vessel segmentation processing on the target coronary angiography image to obtain a target vessel segmentation map includes the following steps: performing vessel segmentation processing on the target coronary angiography image to obtain a candidate vessel segmentation map, and performing noise suppression processing on the candidate vessel segmentation map based on the optimal vessel segmentation map to obtain a target vessel segmentation map. The embodiment of the present application uses the optimal coronary angiography image to suppress the noise of the target coronary angiography image, which is conducive to obtaining a more accurate and reliable target vessel segmentation map, and further can obtain an accurate and reliable main branch vessel length.

[0106] Performing noise suppression processing on the candidate vessel segmentation map based on the optimal vessel segmentation map to obtain a target vessel segmentation map includes the following steps: determining a reference vessel segmentation map, the reference vessel segmentation map is a vessel segmentation map adjacent to the candidate vessel segmentation map and arranged after the candidate vessel segmentation map, determining a vessel intersection area between the candidate vessel segmentation map and the reference vessel segmentation map, mapping the vessel intersection area to the optimal vessel segmentation map, and obtaining a target vessel segmentation map.

[0107] Determining a reference vascular segmentation map includes the following steps: in response to the target coronary angiography image being a coronary angiography image adjacent to the best coronary angiography image, determining the best vascular segmentation map as the reference vascular segmentation map, or, in response to the target coronary angiography image not being a coronary angiography image adjacent to the best coronary angiography image, determining a candidate vascular segmentation map of a coronary angiography image adjacent to the target coronary angiography image and arranged after the target coronary angiography image as the reference vascular segmentation map.

[0108] It can be understood that the mapping method described in the embodiment of the present application refers to extracting the area occupied by the blood vessel intersection area on the optimal blood vessel segmentation map through morphological operations.

[0109] Based on the above-mentioned noise suppression operation, the embodiment of the present application can connect the areas with virtual breaks in the candidate vessel segmentation map through the optimal vessel segmentation map, reduce the adverse effects caused by unclear angiography, and at the same time filter the areas with virtual connections in the candidate vessel segmentation map, so as to obtain a high-quality target vessel segmentation map, which is beneficial to improve the accuracy of calculating the length of the main branch vessel in the subsequent steps.

[0110] In some embodiments, the method for determining the blood vessel flow velocity further comprises the following steps: in response to the sum of pixels of the target blood vessel segmentation image being less than or equal to a preset pixel threshold, or all coronary angiography images of the image segmentation set being in a traversed state, outputting a blood vessel change sequence, wherein the preset pixel threshold is customized by the designer based on engineering experience.

[0111] The target vessel segmentation map that is farther away from the optimal vessel segmentation map contains shorter main vessels. When the sum of the pixels of the target vessel segmentation map is less than or equal to the preset pixel threshold, it means that there will be no main vessels in the subsequent target vessel segmentation map. Therefore, the embodiment of the present application directly outputs the vessel change sequence, that is, outputs all target vessel segmentation maps and the optimal vessel segmentation map. When the sum of the pixels of the target vessel segmentation map is greater than the preset pixel threshold, it means that there are still main vessels in the subsequent target vessel segmentation map. Therefore, the embodiment of the present application needs to continue to determine the target vessel segmentation map.

[0112] Similarly, if all the coronary angiography images in the image segmentation set are in the traversed state, it means that all the coronary angiography images in the image segmentation set have been processed by vascular segmentation, and the embodiment of the present application directly outputs the vascular change sequence. If at least one frame of coronary angiography image in the image segmentation set is in the pending traversal state, it means that there are still coronary angiography images in the image segmentation set that need vascular segmentation processing, and the embodiment of the present application needs to continue to determine the target vascular segmentation map.

[0113] In order to explain in detail the process of noise suppression processing of coronary angiography images in the embodiment of the present application, the embodiment of the present application combines Figure 5 The detailed description is as follows:

[0114] See also Figure 5 After the processing of the above embodiment, the embodiment of the present application obtains: the best vessel segmentation map g1 obtained at the nth time, the candidate vessel segmentation map g2 obtained at the n-1th time, the candidate vessel segmentation map g3 obtained at the n-2th time, the candidate vessel segmentation map g4 obtained at the n-3th time, the candidate vessel segmentation map g5 obtained at the n-4th time, the candidate vessel segmentation map g6 obtained at the n-5th time, and the candidate vessel segmentation map g7 obtained at the n-6th time. Among them, the coronary angiography images of the candidate vessel segmentation maps g2 to g7 are all target coronary angiography images.

[0115] For the candidate vessel segmentation map g2, since the target coronary angiography image of the candidate vessel segmentation map g2 is the coronary angiography image adjacent to the optimal coronary angiography image, the embodiment of the present application selects the optimal vessel segmentation map g1 as the reference vessel segmentation map. The embodiment of the present application determines the vessel intersection region J1 between the candidate vessel segmentation map g2 and the optimal vessel segmentation map g1, maps the vessel intersection region J1 to the optimal vessel segmentation map g1, and obtains the region occupied by the vessel intersection region J1 on the optimal vessel segmentation map g1, and the occupied region is the target vessel segmentation map M1.

[0116] For the candidate vessel segmentation map g3, since the target coronary angiography image of the candidate vessel segmentation map g2 is adjacent to the target coronary angiography image of the candidate vessel segmentation map g3, and the target coronary angiography image of the candidate vessel segmentation map g2 is after the target coronary angiography image of the candidate vessel segmentation map g3, the embodiment of the present application selects the candidate vessel segmentation map g2 as the reference vessel segmentation map. The embodiment of the present application determines the vessel intersection region J2 between the candidate vessel segmentation map g3 and the candidate vessel segmentation map g2, maps the vessel intersection region J2 to the optimal vessel segmentation map g1, and obtains the region occupied by the vessel intersection region J2 on the optimal vessel segmentation map g1, and the occupied region is the target vessel segmentation map M2.

[0117] For the candidate vessel segmentation map g4, since the target coronary angiography image of the candidate vessel segmentation map g3 is adjacent to the target coronary angiography image of the candidate vessel segmentation map g4, and the target coronary angiography image of the candidate vessel segmentation map g3 is after the target coronary angiography image of the candidate vessel segmentation map g4, the embodiment of the present application selects the candidate vessel segmentation map g3 as the reference vessel segmentation map. The embodiment of the present application determines the vessel intersection region J3 between the candidate vessel segmentation map g4 and the candidate vessel segmentation map g3, maps the vessel intersection region J3 to the optimal vessel segmentation map g1, and obtains the region occupied by the vessel intersection region J3 on the optimal vessel segmentation map g1, and the occupied region is the target vessel segmentation map M3.

[0118] For the candidate vessel segmentation map g5, since the target coronary angiography image of the candidate vessel segmentation map g4 is adjacent to the target coronary angiography image of the candidate vessel segmentation map g5, and the target coronary angiography image of the candidate vessel segmentation map g4 is after the target coronary angiography image of the candidate vessel segmentation map g5, the embodiment of the present application selects the candidate vessel segmentation map g4 as the reference vessel segmentation map. The embodiment of the present application determines the vessel intersection region J4 between the candidate vessel segmentation map g5 and the candidate vessel segmentation map g4, maps the vessel intersection region J4 to the optimal vessel segmentation map g1, and obtains the region occupied by the vessel intersection region J4 on the optimal vessel segmentation map g1, and the occupied region is the target vessel segmentation map M4.

[0119] For the candidate vessel segmentation map g6, since the target coronary angiography image of the candidate vessel segmentation map g5 is adjacent to the target coronary angiography image of the candidate vessel segmentation map g6, and the target coronary angiography image of the candidate vessel segmentation map g5 is after the target coronary angiography image of the candidate vessel segmentation map g6, the embodiment of the present application selects the candidate vessel segmentation map g5 as the reference vessel segmentation map. The embodiment of the present application determines the vessel intersection region J5 between the candidate vessel segmentation map g6 and the candidate vessel segmentation map g5, maps the vessel intersection region J5 to the optimal vessel segmentation map g1, and obtains the region occupied by the vessel intersection region J5 on the optimal vessel segmentation map g1, and the occupied region is the target vessel segmentation map M5.

[0120] Since the total pixel sum of the target vessel segmentation map M5 is less than the preset pixel threshold, it means that there will be no main branch vessels in the subsequent target vessel segmentation map. Therefore, the embodiment of the present application directly outputs the vessel change sequence, that is, outputs all target vessel segmentation maps {M1, M2, M3, M4, M5} and the optimal vessel segmentation map g1.

[0121] The embodiment of the present application processes the vascular change sequence based on the skeletonization algorithm to obtain the centerline of the main branch vessel, and calculates the length of the main branch vessel based on the number of pixels contained in the centerline of the main branch vessel and the unit physical length corresponding to each pixel. The embodiment of the present application records the length of the main branch vessel in chronological order to obtain a vascular length change diagram. Please refer to Figure 6 The blood vessel length change graph 61 is provided with a blood vessel length change curve 62, and the blood vessel length change curve is used to indicate the change of the main branch blood vessel length with the conduction time.

[0122] The embodiment of the present application executes step S25 to determine the conduction start time and the conduction end time based on the blood vessel length change graph.

[0123] The conduction start time is the moment when the contrast agent is completely filled at the beginning of the blood vessel and begins to flow forward. The conduction end time is the moment when the contrast agent enters the end of the blood vessel. It can be understood that the units of the conduction start time and the conduction end time are both frames. The difference between the conduction end time and the conduction start time is the TIMI frame number.

[0124] Determining the conduction start time and conduction end time based on the blood vessel length change graph includes the following steps: dividing the blood vessel length change curve into multiple regional segments according to a preset segmentation length, counting the blood vessel length of each regional segment, generating a length histogram based on the blood vessel length of each regional segment, the length histogram is used to represent the distribution of the blood vessel length of each regional segment, and determining the conduction start time and conduction end time based on the length histogram.

[0125] The blood vessel length corresponding to the conduction start time changes quickly, and the value in the length histogram is low. The blood vessel length corresponding to the conduction end time changes slowly, and the value in the length histogram is high. The embodiment of the present application divides the blood vessel length change curve into multiple area segments according to the preset segmentation length, wherein the preset segmentation length is customized by the designer based on engineering experience. For example, please combine Figure 6 , the embodiment of the present application sets the preset segmentation length to 19 pixels, and the range of the blood vessel length change curve is [19, 570]. The embodiment of the present application divides the blood vessel length change curve into multiple regional segments at intervals of 19 pixels, such as the regional segments [19, 38], [38, 57], and so on. Then, the embodiment of the present application counts the blood vessel length in the regional segment [19, 38], the blood vessel length in the regional segment [38, 57], and so on. Finally, the embodiment of the present application analyzes the length histogram to obtain the conduction start time and the conduction end time.

[0126] See also Figure 7 The length histogram 700 shows the blood vessel lengths corresponding to different region segments. The embodiment of the present application analyzes the length histogram 700 to obtain the conduction start time and the conduction end time.

[0127] Determining the conduction start time and the conduction end time based on the length histogram includes the following steps: searching for an area segment that meets the start frame condition among multiple area segments of the length histogram as the start frame segment, searching for an area segment that meets the end frame condition among multiple area segments of the length histogram as the end frame segment, searching for the conduction start time in the blood vessel length change graph based on the segment range corresponding to the start frame segment, and searching for the conduction end time in the blood vessel length change graph based on the segment range corresponding to the end frame segment.

[0128] In some embodiments, searching for an area segment that meets the start frame condition as a start frame segment among multiple area segments of a length histogram includes the following steps: determining that the area segment located on the far left among the multiple area segments of the length histogram is the first target area segment, the first target area segment meets the start frame condition, and the first target area segment is the start frame segment.

[0129] In some other embodiments, searching for an area segment that meets the start frame condition as the start frame segment from multiple area segments of the length histogram includes the following steps: arranging the multiple area segments of the length histogram in order of area length from small to large, obtaining a first area segment set, the first area segment set including a specified number of first area segments, the specified number of first area segments are the specified number of area segments arranged at the front, determining the area segment with the smallest blood vessel length among the specified number of first area segments as the first target area segment, the first target area segment meets the start frame condition, and the first target area segment is the start frame segment.

[0130] For example, the specified number is 3, and the embodiment of the present application selects the three area segments arranged on the leftmost side on the horizontal axis of the length histogram as the first area segments, finds the area segment with the shortest blood vessel length among the three first area segments, and uses the area segment with the shortest blood vessel length as the first target area segment.

[0131] In some embodiments, searching for an area segment that meets an end frame condition among multiple area segments of a length histogram as an end frame segment includes the following steps: determining that the area segment located on the far right among the multiple area segments of the length histogram is a second target area segment, the second target area segment meets the end frame condition, and the second target area segment is an end frame segment.

[0132] In other embodiments, searching for an area segment that meets an end frame condition in a plurality of area segments of a length histogram as an end frame segment includes the following steps: arranging the plurality of area segments of the length histogram in descending order of area length to obtain a second area segment set, the second area segment set including a target number of second area segments, the target number of second area segments being the target number of area segments arranged at the front, determining the area segment with the largest blood vessel length in the target number of second area segments as the second target area segment, the second target area segment meeting the end frame condition, and the second target area segment being the end frame segment.

[0133] For example, the number of targets is 3. In the embodiment of the present application, the three area segments arranged on the far right on the horizontal axis of the length histogram are selected as the second area segments, and the area segment with the largest blood vessel length is found among the three second area segments, and the area segment with the largest blood vessel length is used as the second target area segment.

[0134] The fragment range of a region fragment is constrained by the left endpoint distance and the right endpoint distance.

[0135] In some embodiments, based on the segment range corresponding to the starting frame segment, finding the conduction start time in the blood vessel length change map includes the following steps: determining the left endpoint distance of the segment range corresponding to the starting frame segment as the first target point distance, and finding the time corresponding to the first target point distance in the blood vessel length change map as the conduction start time.

[0136] In other embodiments, based on the segment range corresponding to the starting frame segment, finding the conduction start time in the blood vessel length change map includes the following steps: calculating the average value of the left endpoint distance and the right endpoint distance of the segment range corresponding to the starting frame segment to obtain the first target point distance, and finding the time corresponding to the first target point distance in the blood vessel length change map as the conduction start time.

[0137] In some embodiments, based on the segment range corresponding to the end frame segment, finding the conduction end time in the blood vessel length change graph includes the following steps: determining the left endpoint distance of the segment range corresponding to the end frame segment as the second target point distance, and finding the time corresponding to the second target point distance in the blood vessel length change graph as the conduction end time.

[0138] In other embodiments, based on the segment range corresponding to the end frame segment, finding the conduction end time in the blood vessel length change graph includes the following steps: calculating the average value of the left endpoint distance and the right endpoint distance of the segment range corresponding to the end frame segment to obtain the second target point distance, and finding the time corresponding to the second target point distance in the blood vessel length change graph as the conduction end time.

[0139] In other embodiments, based on the segment range corresponding to the end frame segment, finding the conduction end time in the blood vessel length change map includes the following steps: determining whether the segment range of the end frame segment is the maximum segment range in the second area segment set; if so, determining the left endpoint distance of the segment range of the end frame segment as the second target point distance, and finding the time corresponding to the second target point distance in the blood vessel length change map as the conduction end time; if not, calculating the average value of the left endpoint distance and the right endpoint distance of the segment range corresponding to the end frame segment to obtain the second target point distance, and finding the time corresponding to the second target point distance in the blood vessel length change map as the conduction end time.

[0140] It can be understood that, in most cases, the regional segments of the minimum segment range are all starting frame segments. The embodiment of the present application can take the left endpoint distance of the minimum segment range as the first target point distance, and find the time corresponding to the first target point distance in the blood vessel length change graph as the conduction start time, so that the conduction start time can be determined efficiently.

[0141] It can also be understood that, considering the situation that the regional segment of the maximum segment range is not necessarily the end frame segment, the embodiment of the present application determines whether the segment range of the end frame segment is the maximum segment range in the second regional segment set to determine whether the end segment is the rightmost regional segment. If so, the left endpoint distance of the segment range of the end frame segment is directly taken as the second target point distance. If not, the average value of the left endpoint distance and the right endpoint distance of the segment range corresponding to the end frame segment is calculated to determine the second target point distance, so that the conduction end time can be determined more accurately and reliably.

[0142] Please combine Figure 7 , the embodiment of the present application determines the leftmost region segment 71 in the length histogram 700 as the starting frame segment. Next, the embodiment of the present application determines that the three region segments arranged with the largest region length are all second region segments, and determines the region segment with the largest blood vessel length among the three second region segments as the ending frame segment, such as Figure 7 As shown, the region segment 72 meets the above conditions, and therefore, the region segment 72 is an end frame segment.

[0143] The segment range of the region segment 71 is [19, 38], "19" is the distance to the first target point, and the embodiment of the present application uses "19" as the index. Figure 6 Find out that the time corresponding to "19" is the start time of conduction.

[0144] The segment range [551,570] of the region segment 72 is the maximum segment range. In this embodiment of the application, "551" is directly selected as the second target point distance. Figure 6 Find out that the time corresponding to "551" is the end time of conduction.

[0145] The embodiment of the present application does not need to use a deep learning model to predict the TIMI frame number (i.e., the difference between the conduction start time and the conduction end time). It only identifies the conduction start time and the conduction end time based on the change in the length of the main branch blood vessels, so that the TIMI frame number can be determined simply and quickly.

[0146] The embodiment of the present application performs step S26 to determine the blood vessel flow velocity based on the length of the main branch vessel, the conduction start time and the conduction end time. The embodiment of the present application calculates the blood vessel flow velocity according to the following formula, as shown below:

[0147]

[0148]

[0149] in, and They are the conduction end time and conduction start time, respectively. The length of the main branch vessel, is the frame rate information, is the vascular flow velocity.

[0150] It should be noted that, in each of the above-mentioned embodiments, there is not necessarily a certain order between the above-mentioned steps. A person skilled in the art can understand, based on the description of the embodiments of the present application, that in different embodiments, the above-mentioned steps may have different execution orders, that is, they may be executed in parallel, may be executed interchangeably, and so on.

[0151] As another aspect of the embodiment of the present application, the embodiment of the present application provides a device for determining blood vessel flow rate. The device for determining blood vessel flow rate can be a software module, and the software module includes a number of instructions stored in a memory. The processor can access the memory and call the instructions for execution to complete the method for determining blood vessel flow rate described in the above embodiments.

[0152] In some embodiments, the device for determining the blood vessel flow rate can also be constructed by hardware devices, for example, the device for determining the blood vessel flow rate can be constructed by one or more chips, and each chip can work in coordination with each other to complete the method for determining the blood vessel flow rate described in each of the above embodiments. For another example, the device for determining the blood vessel flow rate can also be constructed by various logic devices, such as a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), a single-chip microcomputer, an ARM (Acorn RISC Machine) or other programmable logic devices, discrete gates or transistor logic, discrete hardware components, or any combination of these components.

[0153] See also Figure 8 The device 800 for determining blood vessel flow velocity includes: an image acquisition module 81, an optimal frame selection module 82, a target image determination module 83, a blood vessel length determination module 84, a TIMI determination module 85 and a flow velocity determination module 86.

[0154] The image acquisition module 81 is used to acquire angiography sequence images, and the angiography sequence images include multiple frames of coronary angiography images. The best frame selection module 82 is used to find the coronary angiography image that meets the best development conditions in the angiography sequence images as the best coronary angiography image. The target image determination module 83 is used to sequentially select the coronary angiography image arranged before the timestamp of the best coronary angiography image in the angiography sequence images as the target coronary angiography image, and the target coronary angiography image and the best coronary angiography image constitute an image segmentation set. The blood vessel length determination module 84 is used to generate a blood vessel length change graph based on the image segmentation set, and the blood vessel length change graph is used to represent the change of the main branch blood vessel length with the conduction time. The TIMI determination module 85 is used to determine the conduction start time and the conduction end time based on the blood vessel length change graph. The flow rate determination module 86 is used to determine the blood vessel flow rate based on the main branch blood vessel length, the conduction start time and the conduction end time.

[0155] The embodiment of the present application does not need to analyze all coronary angiography images, but selects the best coronary angiography image and the target coronary angiography image arranged before the best coronary angiography image for analysis. This can reduce the number of blood vessel segmentations and accurately determine the moment when the contrast agent development effect is best.

[0156] In some embodiments, the vessel length determination module 84 is specifically used to: perform vessel segmentation processing on the optimal coronary angiography image to obtain an optimal vessel segmentation map, perform vessel segmentation processing on the target coronary angiography image to obtain a target vessel segmentation map, the target vessel segmentation map and the optimal vessel segmentation map constitute a vessel change sequence, and generate a vessel length change map based on the vessel change sequence.

[0157] In some embodiments, the vessel length determination module 84 is specifically used to: perform vessel segmentation processing on the target coronary angiography image to obtain a candidate vessel segmentation map, and perform noise suppression processing on the candidate vessel segmentation map based on the optimal vessel segmentation map to obtain a target vessel segmentation map.

[0158] In some embodiments, the vessel length determination module 84 is specifically used to: determine a reference vessel segmentation map, where the reference vessel segmentation map is a vessel segmentation map adjacent to the candidate vessel segmentation map and arranged after the candidate vessel segmentation map, determine a vessel intersection area between the candidate vessel segmentation map and the reference vessel segmentation map, map the vessel intersection area to the optimal vessel segmentation map, and obtain a target vessel segmentation map.

[0159] In some embodiments, the vessel length determination module 84 is specifically used to: in response to the target coronary angiography image being a coronary angiography image adjacent to the optimal coronary angiography image, determine the optimal vessel segmentation map as a reference vessel segmentation map; or, in response to the target coronary angiography image not being a coronary angiography image adjacent to the optimal coronary angiography image, determine a candidate vessel segmentation map of a coronary angiography image adjacent to the target coronary angiography image and arranged after the target coronary angiography image as a reference vessel segmentation map.

[0160] In some embodiments, the vessel length determination module 84 is specifically used to: in response to the sum of pixels of the target vessel segmentation map being less than or equal to a preset pixel threshold, or all coronary angiography images of the image segmentation set being in a traversed state, output the vessel change sequence.

[0161] In some embodiments, the optimal frame selection module 82 is specifically used to: divide all the coronary angiography images of the angiography sequence images into multiple image processing sets, each of the image processing set includes multiple frames of coronary angiography images adjacent to each other, use the image processing set as the input of the pre-trained vascular image processing model, obtain the maximum probability value corresponding to the image processing set, find the maximum probability value with the maximum value among the maximum probability values ​​of all the image processing sets as the target probability value, determine that the coronary angiography image corresponding to the target probability value meets the optimal development condition, and the coronary angiography image corresponding to the target probability value is the optimal coronary angiography image.

[0162] In some embodiments, the blood vessel length change graph is configured with a blood vessel length change curve, and the blood vessel length change curve is used to represent the change of the main branch blood vessel length with the conduction time. The TIMI determination module 85 is specifically used to divide the blood vessel length change curve into multiple regional segments according to a preset segmentation length, count the blood vessel length of each of the regional segments, and generate a length histogram according to the blood vessel length of each of the regional segments. The length histogram is used to represent the distribution of the blood vessel length of each of the regional segments, and the conduction start time and the conduction end time are determined based on the length histogram.

[0163] In some embodiments, the TIMI determination module 85 is specifically used to find a regional segment that meets the start frame condition among the multiple regional segments of the length histogram as the start frame segment, find a regional segment that meets the end frame condition among the multiple regional segments of the length histogram as the end frame segment, find the conduction start time in the blood vessel length change graph based on the segment range corresponding to the start frame segment, and find the conduction end time in the blood vessel length change graph based on the segment range corresponding to the end frame segment.

[0164] It should be noted that the above-mentioned device for determining blood vessel flow velocity can execute the method for determining blood vessel flow velocity provided in the embodiment of the present application, and has the corresponding functional modules and beneficial effects of the execution method. For technical details not fully described in the embodiment of the device for determining blood vessel flow velocity, please refer to the method for determining blood vessel flow velocity provided in the embodiment of the present application.

[0165] See also Fig. 9 , Fig. 9 The schematic diagram of the structure of a computer device provided in an embodiment of the present application is as follows. The computer device 900 includes one or more processors 91 and a memory 92. The memory 92 is connected to the one or more processors 91, for example, via a bus.

[0166] The processor 91 is configured to support the computer device to perform the corresponding functions in the method in the above method embodiment. The processor can be a central processing unit (CPU), a network processor (NP), a hardware chip or any combination thereof. The above hardware chip can be an application specific integrated circuit (ASIC), a programmable logic device (PLD) or a combination thereof. The above PLD can be a complex programmable logic device (CPLD), a field-programmable gate array (FPGA), a generic array logic (GAL) or any combination thereof.

[0167] The memory 92 is used to store program codes, etc. The memory may include a volatile memory (VM), such as a random access memory (RAM); the memory may also include a non-volatile memory (NVM), such as a read-only memory (ROM), a flash memory, a hard disk drive (HDD) or a solid-state drive (SSD); the memory may also include a combination of the above types of memory.

[0168] The memory 92 can be used to store non-volatile software programs, non-volatile computer executable programs and modules, such as program instructions / modules corresponding to the method for determining the blood vessel flow velocity in the embodiment of the present application. The processor executes the various functional applications and data processing of the method for determining the blood vessel flow velocity and the device for determining the blood vessel flow velocity by running the non-volatile software programs, instructions and modules stored in the memory, that is, to realize the functions of the various modules or units of the method for determining the blood vessel flow velocity and the device for determining the blood vessel flow velocity provided in the above method embodiment.

[0169] The memory may include a program storage area and a data storage area, wherein the program storage area may store an operating system and application programs required for at least one function. The data storage area may store data created according to the use of the device for determining the blood vessel flow rate, etc. In some embodiments, the memory may optionally include a memory remotely arranged relative to the processor, and these remote memories may be connected to the device for determining the blood vessel flow rate via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0170] The one or more modules are stored in the memory, and when executed by the one or more processors, the method for determining the vascular flow rate in any of the above-mentioned method embodiments is executed, for example, the method steps described in the above-mentioned method embodiments are executed to realize the functions of the modules described in the above-mentioned device embodiments.

[0171] An embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, wherein the computer program includes program instructions, and when the program instructions are executed by a computer, the computer executes the method described in the above embodiment.

[0172] A person skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes of the embodiments of the above-mentioned methods. The storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc.

[0173] The above disclosure is only the preferred embodiment of the present application, which certainly cannot be used to limit the scope of rights of the present application. Therefore, equivalent changes made according to the claims of the present application are still within the scope covered by the present application.

Claims

1. A method for determining blood vessel flow velocity, characterized in that: include: Acquiring angiography sequence images, wherein the angiography sequence images include multiple frames of coronary angiography images; Finding out a coronary angiography image that meets the best development condition in the angiography sequence images as the best coronary angiography image, wherein finding out a coronary angiography image that meets the best development condition in the angiography sequence images as the best coronary angiography image comprises: dividing all the coronary angiography images in the angiography sequence images into a plurality of image processing sets, each of the image processing sets comprises a plurality of frames of coronary angiography images adjacent to each other, using the image processing set as an input of a pre-trained vascular image processing model, obtaining a maximum probability value corresponding to the image processing set, finding out a maximum probability value of the maximum value among the maximum probability values ​​of all the image processing sets as a target probability value, determining that the coronary angiography image corresponding to the target probability value meets the best development condition, and the coronary angiography image corresponding to the target probability value is the best coronary angiography image; Selecting in sequence from the angiography sequence images a coronary angiography image arranged before the timestamp of the best coronary angiography image as a target coronary angiography image, wherein the target coronary angiography image and the best coronary angiography image form an image segmentation set; A blood vessel length change map is generated based on the image segmentation set, wherein the blood vessel length change map is used to represent the change of the main branch blood vessel length with the conduction time, wherein the blood vessel length change map is generated based on the image segmentation set, including: performing blood vessel segmentation processing on the optimal coronary angiography image to obtain the optimal blood vessel segmentation map, performing blood vessel segmentation processing on the target coronary angiography image to obtain the target blood vessel segmentation map, wherein the target blood vessel segmentation map and the optimal blood vessel segmentation map constitute a blood vessel change sequence, and the blood vessel length change map is generated based on the blood vessel change sequence, wherein the blood vessel segmentation processing on the target coronary angiography image to obtain the target blood vessel segmentation map includes: The image is subjected to a blood vessel segmentation process to obtain a candidate blood vessel segmentation map, and the candidate blood vessel segmentation map is subjected to a noise suppression process based on the optimal blood vessel segmentation map to obtain a target blood vessel segmentation map, wherein the noise suppression process is performed on the candidate blood vessel segmentation map based on the optimal blood vessel segmentation map to obtain the target blood vessel segmentation map, and the process comprises: determining a reference blood vessel segmentation map, wherein the reference blood vessel segmentation map is a blood vessel segmentation map adjacent to the candidate blood vessel segmentation map and arranged after the candidate blood vessel segmentation map, determining a blood vessel intersection region between the candidate blood vessel segmentation map and the reference blood vessel segmentation map, and mapping the blood vessel intersection region to the optimal blood vessel segmentation map to obtain the target blood vessel segmentation map; Determine the conduction start time and conduction end time based on the blood vessel length change graph; The vascular flow velocity is determined based on the main branch vessel length, conduction start time and conduction end time.

2. The determination method according to claim 1, characterized in that: Determining the reference blood vessel segmentation map comprises: In response to the target coronary angiography image being a coronary angiography image adjacent to the best coronary angiography image, determining the best vessel segmentation map as a reference vessel segmentation map; or, In response to the target coronary angiography image not being a coronary angiography image adjacent to the optimal coronary angiography image, a candidate vessel segmentation map of a coronary angiography image adjacent to the target coronary angiography image and arranged after the target coronary angiography image is determined as a reference vessel segmentation map.

3. The determination method according to claim 2, characterized in that: Also includes: In response to the sum of pixels of the target blood vessel segmentation image being less than or equal to a preset pixel threshold, or all coronary angiography images of the image segmentation set being in a traversed state, the blood vessel change sequence is output.

4. The determination method according to any one of claims 1 to 3, characterized in that: The blood vessel length change graph is configured with a blood vessel length change curve, and the blood vessel length change curve is used to represent the change of the main branch blood vessel length with the conduction time. The determination of the conduction start time and the conduction end time based on the blood vessel length change graph includes: According to a preset segmentation length, the blood vessel length variation curve is segmented into a plurality of area segments; Counting the length of blood vessels in each of the regional segments; Generating a length histogram according to the blood vessel length of each of the region segments, wherein the length histogram is used to represent the distribution of the blood vessel lengths of each of the region segments; The conduction start time and the conduction end time are determined based on the length histogram.

5. The determination method according to claim 4, characterized in that: Determining the conduction start time and the conduction end time based on the length histogram comprises: Searching out a region segment satisfying a start frame condition from among a plurality of region segments of the length histogram as a start frame segment; Searching out a region segment satisfying an end frame condition from among a plurality of region segments of the length histogram as an end frame segment; Based on the segment range corresponding to the start frame segment, find out the conduction start time in the blood vessel length change graph; Based on the segment range corresponding to the end frame segment, the conduction end time is found in the blood vessel length change graph.

6. A computer device, characterized in that: It comprises a memory and a processor, wherein the memory is connected to the processor, and the processor is used to execute one or more computer programs stored in the memory. When the processor executes the one or more computer programs, the computer device implements the method for determining the vascular flow velocity as described in any one of claims 1 to 5.

Images