WO2023137948A1 - Processing method and apparatus for analyzing fractional flow reserve on the basis of angiographic image - Google Patents

Abstract

Embodiments of the present invention relates to a processing method and apparatus for analyzing a fractional flow reserve on the basis of an angiographic image. The method comprises: acquiring an angiographic image as a first image; performing object detection and semantic segmentation processing of coronary entry blood vessels and stenotic segment blood vessels on the first image; performing coronary entry diameter analysis to generate a coronary entry diameter; performing stenotic segment blood vessel diameter and stenosis rate analysis to generate a first stenotic segment diameter sequence and a first stenosis rate; taking fractional flow reserve analysis points marked by a user as a targets Pi; confirming to generate corresponding blood vessel branches Ci, and performing blood vessel branch feature extraction to generate a branch feature data sequence and a branch feature data set; inputting the branch feature data set into an artificial neural network model for operation to output a branch operation result set; and outputting branch operation results in the set as fractional flow reserve analysis results of the targets. The present invention can reduce measurement difficulty and improve measurement safety and measurement efficiency.

Description

Processing method and device for analyzing blood flow reserve fraction based on angiography image

This application claims the priority of the Chinese patent application with the application number 202210067548.2 and the title of the invention “Processing Method and Device for Analyzing Blood Flow Reserve Fraction Based on Angiographic Image” submitted to the China Patent Office on January 20, 2022.

technical field

The invention relates to the technical field of data processing, in particular to a processing method and device for analyzing blood flow reserve fraction based on angiography images.

Background technique

Fractional Flow Reserve (FFR) refers to the ratio of the average intracoronary pressure at the distal end of the stenosis to the average aortic pressure at the coronary ostium in the state of maximum coronary artery congestion. Under normal circumstances, the FFR value is obtained by the measurement method of percutaneous coronary intervention (PCI). However, this measurement method is not only complicated to operate, but also an invasive measurement method, which will cause certain physical damage to the measurement object and has certain risks.

Contents of the invention

The purpose of the present invention is to address the defects of the prior art, to provide a processing method, device, electronic equipment and computer-readable storage medium for analyzing blood flow reserve fraction based on angiographic images, to perform quantitative coronary angiography (QCA) analysis on the diameter of the coronary ostium, the diameter change sequence of the stenotic vessel segment, and the stenosis rate of the angiographic image of the coronary artery based on the image target detection and semantic segmentation model; When the FFR measurement point is marked on the above, the artificial neural network (Artificial Neural Network, ANN) model is used to analyze the FFR based on the aforementioned QCA analysis results to obtain the corresponding FFR value. Through the present invention, not only personal injury caused by intrusive measurement can be avoided, but also measurement difficulty can be greatly reduced, and measurement safety and measurement efficiency can be improved.

In order to achieve the above purpose, the first aspect of the embodiment of the present invention provides a processing method for analyzing blood flow reserve fraction based on angiography images, the method comprising:

acquiring an angiographic image as a first image;

Based on the preset image target detection and semantic segmentation model, the first image is subjected to target detection and semantic segmentation processing of coronary inlet vessels and stenotic vessels, so as to obtain a target detection frame A and one or more target detection frames B on the first image; each of the target detection frames A and B includes a section of blood vessel mask image, the detection target type corresponding to the target detection frame A is a coronary artery portal blood vessel type, and the detection target type corresponding to the target detection frame B is a stenotic segment blood vessel type;

Analyzing the coronary artery inlet diameter on the blood vessel mask image of the target detection frame A to generate a corresponding coronary artery inlet diameter; and analyzing the stenotic segment blood vessel diameter and stenosis rate analysis on the blood vessel mask image of each target detection frame B to generate a corresponding first stenosis segment diameter sequence and first stenosis rate;

Taking the blood flow reserve analysis point marked by the user on the first image as the corresponding target point P _i , and taking the user's measurement result of the blood vessel diameter at the target point P _i as the corresponding target point diameter; where 1≤i≤n, n is the total number of measurement points;

Confirm the vascular branch of each target point P _i on the first image to generate a corresponding vascular branch C _i , and perform a vascular branch feature extraction process according to all the first stenotic segment diameter sequences and the first stenosis rate corresponding to each of the vascular branch C _i , the target point diameter of the corresponding target point P _i , and the coronary artery entrance diameter to generate a corresponding branch feature data sequence S _i , and sort all the obtained branch feature data sequences S _i to obtain a branch feature data set D _in (S ₁ ... S _i ... S _n );

Input the branch feature data set D _in (S ₁ ... S _i ... S _n ) into the preset artificial neural network ANN model for operation and output the branch operation result set D _out (U ₁ ... U _i ... U _n ); and output the operation result U _i of each branch in the set as the analysis result of the blood flow reserve fraction of the corresponding target point P _i .

Preferably, the image target detection and semantic segmentation model includes a Mask R-CNN model;

When the image target detection and semantic segmentation model is specifically the Mask R-CNN model, it includes a feature extraction network layer, a region candidate network layer, a region alignment network layer and a region head network layer;

The feature extraction network layer is connected to the region candidate network layer, the region candidate network layer is connected to the region alignment network layer, and the region alignment network layer is connected to the region head network layer;

The feature extraction network layer is specifically composed of a five-level residual network and a corresponding five-level feature pyramid network; the region candidate network layer includes a five-level region candidate network, corresponding to the five-level feature pyramid network; when implementing the five-level residual network, use the residual network ResNet-50 network structure to implement and use this as the backbone network of the feature extraction network layer;

The regional head network layer includes two sub-networks respectively: a target detection branch network and a target segmentation branch network; the target detection branch network is used to output the target detection frame A whose detection target type is a coronary inlet vessel type, and the target detection frame B whose detection target type is a stenotic segment blood vessel type; the target segmentation branch network is used to output the blood vessel mask image of the corresponding coronary inlet vessel in the target detection frame A, and output the corresponding blood vessel mask image of the narrow segment blood vessel in the target detection frame B.

Preferably, the coronary artery entrance diameter analysis is performed on the blood vessel mask image of the target detection frame A to generate a corresponding coronary artery entrance diameter, which specifically includes:

Using the blood vessel mask image of the target detection frame A as the current blood vessel mask image;

Performing blood vessel edge and blood vessel centerline identification on the current blood vessel mask image to generate corresponding current blood vessel edges and current centerline;

The plurality of pixels included in the current centerline are sequentially marked as corresponding first pixel points X _1,j ; wherein, 1≤j≤m ₁ , m ₁ is the total number of pixels in the current centerline; the first first pixel point X _1,j=1 is the blood flow inflow point of the blood vessel corresponding to the current blood vessel mask image, and the last first pixel point

The blood outflow point of the blood vessel corresponding to the current blood vessel mask image;

Through each of the first pixel points X _1,j, do intersection line segments with the current blood vessel edge to obtain a plurality of first intersection line segments; and select the shortest length of the first intersection line segment as the first pixel point blood vessel diameter L _1,j corresponding to the current first pixel point X _1,j ;

Perform mean calculation on all obtained first pixel vessel diameters L _1,j to generate the coronary artery entrance diameter.

Preferably, the analysis of the stenotic vessel diameter and stenosis rate on the blood vessel mask image of each target detection frame B generates a corresponding first stenotic segment diameter sequence and first stenosis rate, specifically including:

Using the blood vessel mask image of the current target detection frame B as the current blood vessel mask image;

Performing blood vessel edge and blood vessel centerline identification on the current blood vessel mask image to generate corresponding current blood vessel edges and current centerline;

The plurality of pixels included in the current centerline are sequentially marked as corresponding second pixel points X _2,h ; wherein, 1≤h≤m ₂ , m ₂ is the total number of pixels in the current centerline; the first second pixel point X _2,h=1 is the blood flow inflow point of the blood vessel corresponding to the current blood vessel mask image, and the last second pixel point

The blood outflow point of the blood vessel corresponding to the current blood vessel mask image;

Through each of the second pixel points X _{2, h} do intersection line segments with the current blood vessel edge to obtain a plurality of second intersection line segments; and select the length of the shortest second intersection line segment therefrom as the second pixel point blood vessel diameter L _2,h corresponding to the current second pixel point X _2,h ;

In addition, according to the second pixel blood vessel diameter L _{2, h=1} and the second pixel blood vessel diameter

The construction can reflect the first second pixel point X _{2, h=1} to the last second pixel point

The linear change function f(h) of the blood vessel linear change relationship,

f(h)=L _2,h=1 +k*(h-1),

And according to the linear change function f(h), calculate the linear change diameter length corresponding to each second pixel point X _2,h to generate the corresponding second pixel point linear diameter L′ _2,h ,

Then according to the second pixel blood vessel diameter L _2,h and the second pixel linear diameter L ^′ _2,h , calculate the second pixel stenosis rate R 2 _,h corresponding to each second pixel X _2,h ,

Using the second pixel point blood vessel diameter L _2,h as the corresponding first stenosis segment diameter, sorting all the first stenosis segment diameters in ascending order according to the second pixel point index h, generating the first stenosis segment diameter sequence corresponding to the current blood vessel mask image; and selecting the maximum value from all the obtained second pixel point stenosis rates R _2,h as the first stenosis rate corresponding to the current blood vessel mask image.

Preferably, the blood vessel branch where each target point P _i is located on the first image is confirmed to generate a corresponding blood vessel branch C _i , and according to all the first stenosis segment diameter sequences and the first stenosis rate corresponding to each of the blood vessel branches C _i , as well as the target point diameter of the corresponding target point P _i , and the coronary artery entrance diameter, a blood vessel branch feature extraction process is performed to generate a corresponding branch feature data sequence S _i , and all obtained branch feature data sequences S _i are sorted to obtain a branch feature data set D _in (S ₁ ... S _i …S _n ), including:

On the first image, the starting position of the blood vessel in the blood vessel mask image of the target detection frame A is used as the coronary artery inlet position; and according to the blood flow direction, the blood flow path from the coronary artery inlet position to each of the target points P _i is used as the blood vessel branch C _i corresponding to each of the target points P _i ;

Classifying the first stenotic segment diameter sequences of all the target detection frames B passed by each of the blood vessel branches _Ci into corresponding first sequence sets; and in each of the first sequence sets, sorting the first stenotic segment diameters of all the first stenotic segment diameter sequences according to the blood flow direction to obtain the corresponding first branch stenotic segment diameter sequences;

From the first stenosis rates of all the target detection frames B passed by each of the blood vessel branches _Ci , select a maximum value as the corresponding maximum stenosis rate of the first branch;

Taking the target diameter of the target point P _i corresponding to each of the blood vessel branches C _i as the corresponding first branch target diameter;

For each of the blood vessel branches C _i , compose the corresponding branch feature data sequence S _i with the coronary inlet diameter and the corresponding first branch stenosis segment diameter sequence, the first branch target diameter and the first branch maximum stenosis rate;

All the branch feature data sequences S _i are sorted in ascending order of the target point index i, so as to obtain the branch feature data set D _in (S ₁ ... S _i ... S _n ).

Preferably, the branch feature data set D _in (S ₁ ... S _i ... S _n ) is input into the preset artificial neural network ANN model for operation, and the branch operation result set D _out (U ₁ ... U _i ... U _n ) is output, specifically including:

Dividing the branch feature data set D _in (S ₁ ... S _i ... S _n ) into n one-dimensional first data segments; each first data segment corresponds to a branch feature data sequence S _i ;

Input each of the first data fragments into the artificial neural network (ANN) model for calculation to obtain the corresponding branch calculation result U _i ;

All the branch operation results U _i are sorted in ascending order of the target point index i, so as to obtain the branch operation result set D _out (U ₁ ... U _i ... U _n ).

Further, the artificial neural network ANN model includes an input layer, one or more hidden layers and an output layer; the input layer includes a plurality of input layer nodes; each of the hidden layers includes a plurality of hidden layer nodes; the output layer includes an output layer node;

The input layer is used to input each segment data of the first data segment to the corresponding input layer node as a corresponding node output value;

Each of the hidden layer nodes of the first hidden layer is fully connected with all the input layer nodes to form a corresponding first fully connected network; the first hidden layer is used to input the node output values of all the input layer nodes in the corresponding first fully connected network into a preset fully connected linear operation function to generate a corresponding first result, and input the first result into a preset activation function to perform operation to generate a corresponding second result, and use the second result as the node output value of the current hidden layer node;

Each of the hidden layer nodes of the next hidden layer is fully connected with all the hidden layer nodes of the previous hidden layer to form a corresponding second fully connected network; the next hidden layer is used to input the node output values of all the hidden layer nodes of the upper hidden layer in the corresponding second fully connected network into a preset fully connected linear operation function to generate a corresponding third result, and input the third result into a preset activation function to perform calculation to generate a corresponding fourth result, and use the fourth result as the node output of the current hidden layer node. value;

The output layer nodes of the output layer are fully connected with all the hidden layer nodes of the last hidden layer to form a corresponding third fully connected network; the output layer is used to input, on the output layer nodes, the node output values of all the hidden layer nodes of the last hidden layer in the corresponding third fully connected network into a preset fully connected linear operation function for operation to generate a corresponding fifth result, and input the fifth result into a preset activation function for operation to generate a corresponding sixth result, and use the sixth result as the branch operation result U _i ;

The activation function used by the hidden layer and the output layer is a ReLU function by default.

The second aspect of the embodiment of the present invention provides a device for implementing the method described in the first aspect above, including: an acquisition module, an image object detection and semantic segmentation module, a quantitative coronary angiography analysis and processing module, a target point processing module, a blood vessel branch processing module, and a blood flow reserve processing module;

The obtaining module is used to obtain an angiographic image as a first image;

The image target detection and semantic segmentation module is used to perform target detection and semantic segmentation processing of coronary inlet vessels and stenotic vessels on the first image based on a preset image target detection and semantic segmentation model, so as to obtain a target detection frame A and one or more target detection frames B on the first image; each of the target detection frames A and B includes a section of blood vessel mask image, the detection target type corresponding to the target detection frame A is a coronary artery blood vessel type, and the detection target type corresponding to the target detection frame B is a stenotic segment blood vessel type;

The quantitative coronary angiography analysis and processing module is used to analyze the coronary artery entrance diameter on the blood vessel mask image of the target detection frame A to generate a corresponding coronary artery entrance diameter; and analyze the stenotic segment blood vessel diameter and stenosis rate analysis on the blood vessel mask image of each target detection frame B to generate a corresponding first stenosis segment diameter sequence and first stenosis rate;

The target point processing module is used to use the blood flow reserve fraction analysis point marked by the user on the first image as the corresponding target point P _i , and use the user's blood vessel diameter measurement result at the target point P _i as the corresponding target point diameter; where 1≤i≤n, n is the total number of measurement points;

The blood vessel branch processing module is used to confirm the blood vessel branch of each target point P _i on the first image to generate a corresponding blood vessel branch C _i , and perform blood vessel branch feature extraction processing according to all the first stenosis segment diameter sequences and the first stenosis rate corresponding to each of the blood vessel branch C _i , the target point diameter of the corresponding target point _Pi , and the coronary artery entrance diameter to generate a corresponding branch feature data sequence S _i , and sort all the obtained branch feature data sequences S _i to obtain a branch feature data set D _in (S ₁ ... S _i ... S _n );

The blood flow reserve fraction processing module is used to input the branch feature data set D _in (S ₁ ... S _i ... S _n ) into the preset artificial neural network ANN model for operation and output a branch operation result set D _out (U ₁ ... U _i ... U _n ); and output the operation result U _i of each branch in the set as the analysis result of the corresponding target point P _i .

The third aspect of the embodiment of the present invention provides an electronic device, including: a memory, a processor, and a transceiver;

The processor is configured to be coupled with the memory, read and execute instructions in the memory, so as to implement the method steps described in the first aspect above;

The transceiver is coupled to the processor, and the processor controls the transceiver to send and receive messages.

The fourth aspect of the embodiments of the present invention provides a computer-readable storage medium, the computer-readable storage medium stores computer instructions, and when the computer instructions are executed by a computer, the computer executes the instructions of the method described in the first aspect above.

An embodiment of the present invention provides a processing method, device, electronic device, and computer-readable storage medium for analyzing blood flow reserve fraction based on angiographic images. Based on the image target detection and semantic segmentation model, the angiographic image of the coronary artery is used to identify the coronary ostium vessel segment and the stenotic vessel segment, and perform QCA analysis on the diameter of the coronary ostium vessel segment, the diameter change sequence of the stenotic vessel segment, and the stenosis rate; FR value. The invention not only avoids personal injury caused by intrusive measurement, but also greatly reduces measurement difficulty, and improves measurement safety and measurement efficiency.

Description of drawings

FIG. 1 is a schematic diagram of a processing method for analyzing blood flow reserve fraction based on angiographic images provided by Embodiment 1 of the present invention;

FIG. 2 is a block diagram of a processing device for analyzing blood flow reserve fraction based on angiographic images provided by Embodiment 2 of the present invention;

FIG. 3 is a schematic structural diagram of an electronic device provided by Embodiment 3 of the present invention.

Detailed ways

In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings. Apparently, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present invention.

Embodiment 1 of the present invention provides a processing method for analyzing blood flow reserve based on angiographic images, as shown in FIG. 1 , a schematic diagram of a processing method for analyzing blood flow reserve based on angiographic images provided in Embodiment 1 of the present invention. This method mainly includes the following steps:

Step 1, acquiring an angiographic image as a first image.

Here, the coronary angiography technique is to inject a contrast agent into the blood vessel of the measurement object and take an image of the process of the contrast agent passing through the coronary artery under X-ray; the image data obtained by the coronary angiography technique is a coronary angiography image; the angiography image in the current step defaults to a coronary angiography image in which the contrast agent filling effect is more obvious in a single coronary angiography process.

Step 2. Based on the preset image target detection and semantic segmentation model, the first image is subjected to target detection and semantic segmentation processing of coronary inlet vessels and stenotic vessels, so as to obtain a target detection frame A and one or more target detection frames B on the first image;

Wherein, the target detection frame A and B both include a section of blood vessel mask image, the detection target type corresponding to the target detection frame A is the coronary inlet blood vessel type, and the detection target type corresponding to the target detection frame B is the stenotic segment blood vessel type;

There are many ways to implement the image target detection and semantic segmentation model, one of which is based on the neural network architecture of the Mask R-CNN model; when the image target detection and semantic segmentation model is specifically the Mask R-CNN model, its neural network structure can refer to the article "Mask R-CNN" published by the authors Kaiming He, Georgia Gkioxari, Piotr Doll'ar and Ross Girshick, including: feature extraction network layer, region candidate network (Region Propos al Network, RPN) layer, region alignment (Region Of Interest Align, ROI Align) network layer and region head (ROI HEAD) network layer; the feature extraction network layer is connected to the region candidate network layer, the region candidate network layer is connected to the region alignment network layer, and the region alignment network layer is connected to the region head network layer;

The feature extraction network layer of the embodiment of the present invention is specifically composed of a five-level residual network (Residual Network, ResNet) and a corresponding five-level feature pyramid network (Feature Pyramid Networks, FPN); the region candidate network layer includes a five-level region candidate network, corresponding to the five-level feature pyramid network; when implementing a five-level residual network, the embodiment of the present invention uses the residual network ResNet-50 network structure to implement and use this as the backbone network of the feature extraction network layer;

The regional head network layer in the embodiment of the present invention includes two sub-networks: a target detection branch network and a target segmentation branch network; wherein, the target detection branch network is used to output a target detection frame A whose detection target type is a coronary inlet vessel type, and a target detection frame B whose detection target type is a stenotic blood vessel type; the target segmentation branch network is used to output a blood vessel mask (mask) image of a corresponding coronary inlet vessel in the target detection frame A, and output a blood vessel mask image of a corresponding stenotic blood vessel in the target detection frame B.

Step 3: Analyzing the coronary artery entrance diameter on the blood vessel mask image of the target detection frame A to generate the corresponding coronary artery entrance diameter; and analyzing the stenotic vessel diameter and stenosis rate analysis on the blood vessel mask image of each target detection frame B to generate a corresponding first stenosis segment diameter sequence and first stenosis rate;

It specifically includes: step 31, analyzing the coronary artery entrance diameter on the blood vessel mask image of the target detection frame A to generate a corresponding coronary artery entrance diameter;

It specifically includes: step 311, using the blood vessel mask image of the target detection frame A as the current blood vessel mask image;

Step 312, performing vessel edge and vessel centerline recognition on the current vessel mask image to generate corresponding current vessel edge and current centerline;

Here, when the blood vessel edge and the blood vessel centerline are identified on the current blood vessel mask image, all the image information in the target detection frame A where the current blood vessel mask image is located is firstly used as the first detection frame image; then, the first detection frame image is binarized to generate the first binary image; wherein, there are only two types of pixel values for all pixels on the first binary image: the foreground pixel value of the foreground pixel point, and the background pixel value of the background pixel point. The foreground pixel value and the background pixel value can be defined by themselves during specific implementation;

Then, traverse all the pixels covered by the current blood vessel mask image on the first binary image, if there is at least one background pixel in the four-neighborhood pixel range of the currently traversed pixel during traversal, then use the currently traversed pixel as the blood vessel edge pixel, and after completing the traversal, connect all the blood vessel edge pixels clockwise or counterclockwise to obtain the closed blood vessel edge curve of the current blood vessel mask image, that is, the current blood vessel edge;

In addition, under the premise of not changing the topological properties of the blood vessel image (connectivity of the blood vessel), based on the topology thinning method, the centerline of the current blood vessel mask image on the first binary image is extracted to generate the current centerline, specifically: create a connectivity judgment rule including multiple sub-rules for the connectivity of the blood vessel in advance, each sub-rule corresponds to one or more 3×3 pixel value template matrices, and each pixel value template matrix corresponds to a pixel arrangement structure that maintains connectivity; then traverse all the pixels covered by the current blood vessel mask image on the first binary image; The pixel values of the pixels in the eight domains of the current traversing pixel are extracted and combined with the pixel values of the currently traversing pixel to form a 3×3 pixel value matrix, which is recorded as the first matrix, and then one or more 3×3 pixel value template matrices of each sub-rule in the connectivity judgment rule are used to compare with the first matrix in turn; when comparing, if the first matrix completely matches any of the pixel value template matrices, it means that the currently traversing pixel is a key pixel to maintain the connectivity of neighboring pixels and cannot be deleted, and jumps to the next pixel for traversal; comparison , if the first matrix does not match the pixel value template matrix of all the sub-rules in the connectivity judgment rule, it means that the current traversal pixel is not a key pixel to maintain the connectivity of neighboring pixels and can be deleted, correspondingly marked as a deletable pixel; after completing a traversal of all the pixels covered by the current vascular mask image, all the pixels marked as deletable pixels are converted into background pixels, thus realizing a topological refinement of the current vascular mask image without destroying the topological properties (vascular connectivity) After converting all the deletable pixels into background pixels, continue to traverse all the pixels covered by the topologically thinned current blood vessel mask image according to the above steps, and judge whether each traversal pixel is a key pixel according to the pixel value template matrix of the connectivity judgment rule during traversal, and whether it needs to be reserved or marked as a deletable pixel, and after the end of the current traversal, all the pixels marked as deletable pixels are converted into background pixels; and so on until the end of the current traversal The number of pixels marked as deletable is 0, indicating that the current blood vessel mask image cannot be further topologically refined. At this time, the remaining pixels in the current blood vessel mask image are connected in sequence according to the blood flow direction, and the connected curve is the current center line;

Step 313, mark the plurality of pixel points included in the current central line as corresponding first pixel points X _1,j in sequence;

Among them, 1≤j≤m ₁ , m ₁ is the total number of pixels of the current centerline; the first first pixel point X _1,j=1 is the blood flow inflow point of the blood vessel corresponding to the current blood vessel mask image, and the last first pixel point X _1,j=m1 is the blood flow outflow point of the blood vessel corresponding to the current blood vessel mask image;

Step 314, through each first pixel point X _1,j, make the intersection line segment with the edge of the current blood vessel to obtain a plurality of first intersection line segments; and select the length of the shortest first intersection line segment as the first pixel point blood vessel diameter L _1,j corresponding to the current first pixel point X _1,j ;

Here, according to each first pixel point X _1,jThe arrangement structure relationship of the pixels in its eight neighborhoods (upper left, upper, upper right, right, lower right, lower, lower left, left), each first pixel point X _1,jMake four straight lines: the first straight line (upper left-X _1,j- bottom left), second straight line (above - X _1,j-bottom), third straight line (upper right-X _1,j-bottom left) and the fourth straight line (right-X _1,j-left); extract the first, second, third and fourth straight lines and the intersection line segment of the current blood vessel edge as the corresponding first intersection line segment 1, 2, 3, 4; from the first intersection line segment 1, 2, 3, 4, extract the shortest as the first pixel point blood vessel diameter L _1,j;

Step 315, performing mean value calculation on all obtained first pixel vessel diameters L _1,j to generate a coronary artery entrance diameter;

Here, using the mean calculation method can reduce the calculation error in the image processing process;

Step 32, analyzing the vessel diameter and stenosis rate of the stenosis segment on the blood vessel mask image of each target detection frame B to generate a corresponding first stenosis segment diameter sequence and first stenosis rate;

It specifically includes: step 321, using the blood vessel mask image of the current target detection frame B as the current blood vessel mask image;

Step 322, performing vessel edge and vessel centerline recognition on the current vessel mask image to generate corresponding current vessel edge and current centerline;

Step 323, mark the plurality of pixel points included in the current central line as corresponding second pixel points X _2,h in sequence;

Among them, 1≤h≤m ₂ , m ₂ is the total number of pixels of the current center line; the first second pixel point X _{2, h=1} is the blood flow inflow point of the blood vessel corresponding to the current blood vessel mask image, and the last second pixel point

The blood outflow point of the blood vessel corresponding to the current blood vessel mask image;

Step 324, through each second pixel point X _{2, h,} make a plurality of second intersecting line segments with the current blood vessel edge; and select the length of the shortest second intersecting line segment as the second pixel point blood vessel diameter L 2 _{, h} corresponding to the current second pixel point X _{2, h} ;

Here, it is similar to step 314, and no further details are given here;

Step 325, according to the second pixel blood vessel diameter L _{2, h=1} and the second pixel blood vessel diameter

The construction can reflect the first second pixel point X _{2, h=1} to the last second pixel point

The linear change function f(h) of the blood vessel linear change relationship,

f(h)=L _2,h=1 +k*(h-1),

Here, the embodiment of the present invention defaults that the diameter of the coronary artery from the coronary artery entrance to the end point of each branch should be a linear reduction process under normal conditions without stenosis of the middle segment of the blood vessel caused by the lesion. The above-mentioned linear change function f(h) is a function for simulating this linear reduction process;

Step 326, according to the linear change function f(h), calculate the linear change diameter length corresponding to each second pixel point X _2,h to generate the corresponding second pixel point linear diameter L′ _2,h ,

Here, the second pixel point linear diameter L′ _2,h is the normal blood vessel diameter corresponding to each second pixel point X _2,h under normal circumstances based on the above-mentioned linear change function f(h);

Step 327, according to the second pixel blood vessel diameter L _2,h and the second pixel linear diameter L′ _2,h , calculate the second pixel stenosis rate R _2,h corresponding to each second pixel X _2,h ,

Here, the second pixel point blood vessel diameter L _2,h is the blood vessel diameter under actual conditions, and the second pixel point linear diameter L′ _2,h is the theoretical blood vessel diameter under simulated normal conditions,

The ratio of the actual diameter to the theoretical diameter can reflect the degree of blood flow through the second pixel point X _2,h , naturally

It reflects the degree of blood flow obstruction of the second pixel point X _2,h , that is, the second pixel point stenosis rate R _2,h corresponding to the second pixel point X _2,h ;

Step 328, using the second pixel point blood vessel diameter L _2,h as the corresponding first stenosis segment diameter, sorting all the first stenosis segment diameters in ascending order according to the second pixel point index h, generating a sequence of first stenosis segment diameters corresponding to the current blood vessel mask image; and selecting the maximum value from all obtained second pixel point stenosis ratios R _2,h as the first stenosis rate corresponding to the current blood vessel mask image.

Here, the first series of stenotic segment diameters reflects the diameters of blood vessels at various points on the blood vessel corresponding to the stenotic segment, and the first stenosis rate is the maximum stenosis rate on the blood vessel corresponding to the stenotic segment.

It should be noted that when calculating the first stenosis rate corresponding to the current blood vessel mask image, the embodiment of the present invention supports a variety of processing methods; one of them can be seen from the above steps 327-328, first calculate the complete stenosis rate sequence, and then select the maximum value as the first stenosis rate; the other is to first select the minimum value from all second pixel blood vessel diameters L _2,h as the second pixel minimum blood vessel diameter L _min , and substitute the pixel index y corresponding to L _min into the linear change function f(h=y) to calculate and obtain the pixel index The linear reference diameter L′ _{2,h=y corresponding to y} , and then substitute L _min and L′ _2,h=y into

Perform calculation to obtain a calculation result R, and then output R as the first stenosis rate. The feature of the former of these two processing methods is that the stenosis rate of each position on the centerline of the stenosis can be obtained while obtaining the first stenosis rate, and the feature of the latter is that the calculation method is simpler and faster because it does not need to output the stenosis rate of each position.

The above step 3 is the QCA analysis and processing process completed by using the image processing method, through which the analysis results of the coronary entrance diameter and all stenotic segment diameter sequences and stenosis ratios on the first image can be obtained. On this basis, the embodiment of the present invention can further measure the FFR value of the measurement point marked on the blood vessel in any stenotic segment on the first image through the subsequent steps 4-6.

Step 4, taking the fractional blood flow reserve analysis point marked by the user on the first image as the corresponding target point P _i , and taking the user's blood vessel diameter measurement result at the target point P _i as the corresponding target point diameter;

Among them, i is the index of the target point, 1≤i≤n, and n is the total number of measurement points.

Here, the target point P _i is the blood flow reserve fraction measurement _point of the stenotic segment marked by the user on the first image. The measurement point is a position point on a certain segment of the blood vessel on the first image. Normally, the user will mark the target point _{P i} on the blood flow branch of the stenotic segment of the blood vessel that is a certain distance from the last stenotic segment of the branch. The measurement result is also the corresponding _target spot diameter.

Step 5: Confirm the vascular branch of each target point P _i on the first image to generate the corresponding vascular branch C _i , and perform vascular branch feature extraction processing according to all first stenotic segment diameter sequences and first stenosis ratios corresponding to each vascular branch C _i , as well as the target point diameter of the corresponding target point P _i , and the coronary artery entrance diameter to generate a corresponding branch feature data sequence S _i , and sort all the obtained branch feature data sequences S _i to obtain a branch feature data set D _in (S ₁ ... S _i ... S _n );

It specifically includes: step 51, on the first image, taking the initial position of the blood vessel in the blood vessel mask image of the target detection frame A as the coronary artery inlet position; and according to the blood flow direction, taking the blood flow path from the coronary artery inlet position to each target point P _i as the blood vessel branch C _i corresponding to each target point P _i ;

Here, the blood vessel branch C _i refers to a single blood flow path from the coronary inlet position to each target point P _i position;

Step 52, classify the first stenotic segment diameter sequences of all target detection frames B passed by each blood vessel branch _Ci into the corresponding first sequence set; and in each first sequence set, sort the first stenotic segment diameters of all the first stenotic segment diameter sequences according to the blood flow direction to obtain the corresponding first branch stenotic segment diameter sequence;

Here, if from the coronary entrance position to the current target point P _iThere is only one stenotic vessel segment in the single blood flow path at the position, then the vessel branch C _iOnly one target detection frame B is passed, and the corresponding first sequence set only includes the first stenosis diameter sequence of the unique target detection frame B, and the first branch stenosis diameter sequence is also the first stenosis diameter sequence of the unique target detection frame B; if from the coronary artery entrance position to the current target point P _iA single blood flow path at the position passes through multiple narrow vessel segments, then the vessel branch C _iAfter passing through multiple target detection frames B, the corresponding first sequence set includes the first stenotic segment diameter sequences of multiple target detection frames B. Sorting these multiple first stenotic segment diameter sequences according to the blood flow direction realizes the full sorting process in this step. The obtained first branch stenotic segment diameter sequence includes the current vessel branch C _iFrom the position of the coronary artery entrance to the corresponding target point P _iVessel diameter change parameters of all stenotic vessels on the blood flow path at the position;

It should be noted that, after obtaining the diameter sequence of the narrow section of the first branch, in order to facilitate the subsequent model calculation, the embodiment of the present invention will also perform data shaping on it to ensure that its data length is consistent with the required length of the model;

Step 53, from the first stenosis rates of all target detection frames B passed by each blood vessel branch _Ci , select the maximum value as the corresponding maximum stenosis rate of the first branch;

Here, if the number of target detection frames B passed by the blood vessel branch C _i is unique, the maximum stenosis rate of the first branch is the maximum stenosis rate of the unique target detection frame B, that is, the first stenosis rate; if the number of target detection frames B passed by the blood vessel branch C _i is not unique, then the maximum stenosis rate of the first branch should be the maximum value among the first stenosis rates of multiple target detection frames B, that is, the maximum stenosis rate on a single blood flow path corresponding to the blood vessel branch C _i ;

Step 54, using the target diameter of the target point P _i corresponding to each blood vessel branch C _i as the corresponding first branch target diameter;

Here, the diameter of the first branch target point corresponding to each vascular branch C _i is actually the cross-sectional diameter of the blood vessel at the position of the target point P _i at the end of the branch;

Step 55, for each vascular branch C _i , compose the corresponding branch feature data sequence S _i with the diameter of the coronary artery inlet, the corresponding first branch stenosis segment diameter sequence, the first branch target point diameter and the first branch maximum stenosis rate;

Here, there are multiple ways of data combination of the branch feature data sequence S _i composed of the coronary artery inlet diameter, the first branch stenosis segment diameter sequence, the first branch target point diameter and the first branch maximum stenosis rate, one of which is: branch feature data sequence S _i =coronary artery inlet diameter+the first branch stenosis segment diameter sequence+the first branch target point diameter+the first branch maximum stenosis rate;

Step 56, sort all the branch feature data sequences S _i in ascending order of the target point index i, so as to obtain the branch feature data set D _in (S ₁ ... S _i ... S _n ).

Step 6, input the branch feature data set D _in (S ₁ ... S _i ... S _n ) into the preset artificial neural network ANN model for operation and output the branch operation result set D _out (U ₁ ... U _i ... U _n ); and output the branch operation result U _i in the set as the blood flow reserve fraction analysis result of the corresponding target point P _i ;

It specifically includes: step 61, inputting the branch feature data set D _in (S ₁ ... S _i ... S _n ) into the preset artificial neural network ANN model for operation and outputting the branch operation result set D _out (U ₁ ... U _i ... U _n );

It specifically includes: step 611, dividing the branch feature data set D _in (S ₁ ... S _i ... S _n ) into n one-dimensional first data segments; each first data segment corresponds to a branch feature data sequence S _i ;

Step 612, input each first data segment into the artificial neural network ANN model for calculation to obtain the corresponding branch calculation result U _i ;

Wherein, the artificial neural network ANN model includes an input layer, one or more hidden layers and an output layer; the input layer includes a plurality of input layer nodes; each hidden layer includes a plurality of hidden layer nodes; the output layer includes an output layer node;

When the artificial neural network ANN model operates on the first input data segment:

The input layer is used for inputting each segment data of the first data segment into a corresponding input layer node as a corresponding node output value;

Each hidden layer node of the first hidden layer is fully connected with all input layer nodes to form a corresponding first fully connected network; the first hidden layer is used to input the node output values of all input layer nodes in the corresponding first fully connected network into a preset fully connected linear operation function to generate a corresponding first result, and input the first result to a preset activation function to generate a corresponding second result, and use the second result as the node output value of the current hidden layer node;

Each hidden layer node of the next hidden layer is fully connected with all hidden layer nodes of the previous hidden layer to form a corresponding second fully connected network; the hidden layer of the next layer is used to input the node output values of all hidden layer nodes of the previous hidden layer in the corresponding second fully connected network into a preset fully connected linear operation function to generate a corresponding third result, and input the third result into a preset activation function to perform operation to generate a corresponding fourth result, and use the fourth result as the node output value of the current hidden layer node;

The output layer nodes of the output layer are fully connected with all hidden layer nodes of the last hidden layer to form a corresponding third fully connected network; the output layer is used to input the node output values of all hidden layer nodes of the last hidden layer in the corresponding third fully connected network on the output layer nodes into a preset fully connected linear operation function for operation to generate a corresponding fifth result, and input the fifth result into a preset activation function for operation to generate a corresponding sixth result, and use the sixth result as the branch operation result U _i ;

In the above operation process, the activation function used by the hidden layer and the output layer defaults to the Rectified Linear Units (ReLU) function;

Step 613, sort all the branch operation results U _i according to the order of the target point index i from small to large, so as to obtain the branch operation result set D _out (U ₁ ... U _i ... U _n );

Step 62 , output the operation result U _i of each branch in the set as the fractional blood flow reserve analysis result of the corresponding target point P _i .

2 is a module structure diagram of a processing device for analyzing blood flow reserve fraction based on angiographic images provided by Embodiment 2 of the present invention. The device may be a terminal device or a server implementing the method of the embodiment of the present invention, or a device connected to the above-mentioned terminal device or server to realize the method of the embodiment of the present invention. For example, the device may be a device or a chip system of the above-mentioned terminal device or server. As shown in FIG. 2 , the device includes: an acquisition module 201 , an image target detection and semantic segmentation module 202 , a quantitative coronary angiography analysis and processing module 203 , a target point processing module 204 , a vessel branch processing module 205 and a blood flow reserve processing module 206 .

The obtaining module 201 is used to obtain an angiographic image as a first image.

The image target detection and semantic segmentation module 202 is used to perform target detection and semantic segmentation on the first image based on the preset image target detection and semantic segmentation model, so as to obtain a target detection frame A and one or more target detection frames B on the first image; both target detection frames A and B include a section of blood vessel mask image, and the detection target type corresponding to the target detection frame A is the coronary entrance blood vessel type, and the detection target type corresponding to the target detection frame B is the stenotic segment blood vessel type.

The quantitative coronary angiography analysis and processing module 203 is used to analyze the diameter of the coronary artery entrance on the blood vessel mask image of the target detection frame A to generate the corresponding coronary artery entrance diameter; and analyze the vessel diameter and stenosis rate of each target detection frame B to generate a corresponding first stenosis segment diameter sequence and first stenosis rate.

The target point processing module 204 is used to use the blood flow reserve fraction analysis point marked by the user on the first image as the corresponding target point P _i , and use the user's blood vessel diameter measurement result at the target point P _i as the corresponding target point diameter; where i is the target point index, 1≤i≤n, and n is the total number of measurement points.

The blood vessel branch processing module 205 is used to confirm the blood vessel branch of each target point P _i on the first image to generate the corresponding blood vessel branch C _i , and perform blood vessel branch feature extraction processing according to all first stenosis segment diameter sequences and first stenosis rates corresponding to each blood vessel branch C _i , the target point diameter of the corresponding target point P _i , and the diameter of the coronary artery entrance to generate a corresponding branch feature data sequence S _i , and sort all the obtained branch feature data sequences S _i to obtain a branch feature data set D _in (S ₁ ... S _i ... S _n ).

The fractional blood flow reserve processing module 206 is used to input the branch feature data set D _in (S ₁ ... S _i ... S _n ) into the preset artificial neural network ANN model for calculation and output the branch operation result set D _out (U ₁ ... U _i ... U _n ); and output the operation result U _i of each branch in the set as the analysis result of the blood flow reserve fraction of the corresponding target point P _i .

The embodiment of the present invention provides a processing device for analyzing blood flow reserve fraction based on angiography images, which can execute the method steps in the above method embodiments, and its implementation principle and technical effect are similar, and will not be repeated here.

It should be noted that it should be understood that the division of each module of the above device is only a division of logical functions, and may be fully or partially integrated into one physical entity or physically separated during actual implementation. And these modules can all be implemented in the form of calling software through processing elements; they can also be implemented in the form of hardware; some modules can also be implemented in the form of calling software through processing elements, and some modules can be implemented in the form of hardware. For example, the acquisition module may be a separate processing element, or may be integrated into a certain chip of the above-mentioned device. In addition, it may also be stored in the memory of the above-mentioned device in the form of program code, and a certain processing element of the above-mentioned device calls and executes the function of the above determination module. The implementation of other modules is similar. In addition, all or part of these modules can be integrated together, and can also be implemented independently. The processing element described here may be an integrated circuit with signal processing capability. In the implementation process, each step of the above method or each module above can be completed by an integrated logic circuit of hardware in the processor element or an instruction in the form of software.

For example, the above modules may be one or more integrated circuits configured to implement the above method, for example: one or more specific integrated circuits (Application Specific Integrated Circuit, ASIC), or, one or more digital signal processors (Digital Signal Processor, DSP), or, one or more field programmable gate arrays (Field Programmable Gate Array, FPGA) and the like. For another example, when one of the above modules is implemented in the form of a processing element scheduling program code, the processing element may be a general-purpose processor, such as a central processing unit (Central Processing Unit, CPU) or other processors that can call program codes. For another example, these modules can be integrated together and implemented in the form of a System-on-a-chip (SOC).

In the above embodiments, all or part of them may be implemented by software, hardware, firmware or any combination thereof. When implemented using software, it may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on the computer, all or part of the processes or functions described in accordance with the embodiments of the present invention will be generated. The above-mentioned computers may be general-purpose computers, special-purpose computers, computer networks, or other programmable devices. The above-mentioned computer instructions may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, the above-mentioned computer instructions may be transmitted from one website, computer, server or data center to another website, computer, server or data center through wired (such as coaxial cable, optical fiber, digital subscriber line (Digital Subscriber Line, DSL)) or wireless (such as infrared, wireless, bluetooth, microwave, etc.). The above-mentioned computer-readable storage medium may be any available medium that can be accessed by a computer, or a data storage device such as a server or a data center integrated with one or more available media. The above-mentioned usable medium may be a magnetic medium (for example, a floppy disk, a hard disk, or a magnetic tape), an optical medium (for example, DVD), or a semiconductor medium (for example, a solid state disk (solid state disk, SSD)) and the like.

FIG. 3 is a schematic structural diagram of an electronic device provided by Embodiment 3 of the present invention. The electronic device may be the aforementioned terminal device or server, or may be a terminal device or server connected to the aforementioned terminal device or server to implement the method of the embodiment of the present invention. As shown in FIG. 3 , the electronic device may include: a processor 301 (such as a CPU), a memory 302 , and a transceiver 303 ; Various instructions may be stored in the memory 302 for completing various processing functions and realizing the methods and processing procedures provided in the above-mentioned embodiments of the present invention. Preferably, the electronic device involved in this embodiment of the present invention further includes: a power supply 304 , a system bus 305 and a communication port 306 . The system bus 305 is used to realize the communication connection among the components. The above-mentioned communication port 306 is used for connection and communication between the electronic device and other peripheral devices.

The system bus mentioned in FIG. 3 may be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus or the like. The system bus can be divided into address bus, data bus, control bus and so on. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus. The communication interface is used to realize the communication between the database access device and other devices (such as client, read-write library and read-only library). The memory may include random access memory (Random Access Memory, RAM), and may also include non-volatile memory (Non-Volatile Memory), such as at least one disk memory.

The above-mentioned processor can be a general-purpose processor, including a central processing unit CPU, a network processor (Network Processor, NP), etc.; it can also be a digital signal processor DSP, an application-specific integrated circuit ASIC, a field programmable gate array FPGA or other programmable logic devices, discrete gates or transistor logic devices, and discrete hardware components.

It should be noted that the embodiments of the present invention also provide a computer-readable storage medium, and instructions are stored in the storage medium, and when the storage medium is run on a computer, the computer executes the methods and processing procedures provided in the above-mentioned embodiments.

The embodiment of the present invention also provides a chip for running instructions, and the chip is used for executing the method and the processing procedure provided in the foregoing embodiments.

An embodiment of the present invention provides a processing method, device, electronic device, and computer-readable storage medium for analyzing blood flow reserve fraction based on angiographic images. Based on the image target detection and semantic segmentation model, the angiographic images of coronary vessels are used to identify the coronary ostium vessel segment and the stenotic vessel segment, and perform QCA analysis on the diameter of the identified coronary ostium vessel segment, the diameter change sequence of the stenotic vessel segment, and the stenosis rate; when the user marks the FFR measurement point on the image, use the ANN model to analyze the FFR based on the aforementioned QCA analysis results to obtain the corresponding FFR value . The invention not only avoids personal injury caused by intrusive measurement, but also greatly reduces measurement difficulty, and improves measurement safety and measurement efficiency.

Professionals should further realize that the units and algorithm steps of the examples described in the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described according to their functions in the above description. Whether these functions are executed by hardware or software depends on the specific application and design constraints of the technical solution. Skilled artisans may use different methods to implement the described functions for each specific application, but such implementation should not be regarded as exceeding the scope of the present invention.

The steps of the methods or algorithms described in connection with the embodiments disclosed herein may be implemented by hardware, software modules executed by a processor, or a combination of both. The software module can be placed in random access memory (RAM), internal memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the technical field.

The above-mentioned specific implementation manners further describe the purpose, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above-mentioned descriptions are only specific implementation modes of the present invention, and are not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims (10)

1. A processing method for analyzing blood flow reserve fraction based on angiographic images, characterized in that the method comprises:

   acquiring an angiographic image as a first image;

   Based on the preset image target detection and semantic segmentation model, the first image is subjected to target detection and semantic segmentation processing of coronary inlet vessels and stenotic vessels, so as to obtain a target detection frame A and one or more target detection frames B on the first image; each of the target detection frames A and B includes a section of blood vessel mask image, the detection target type corresponding to the target detection frame A is a coronary artery portal blood vessel type, and the detection target type corresponding to the target detection frame B is a stenotic segment blood vessel type;

   Analyzing the coronary artery inlet diameter on the blood vessel mask image of the target detection frame A to generate a corresponding coronary artery inlet diameter; and analyzing the stenotic segment blood vessel diameter and stenosis rate analysis on the blood vessel mask image of each target detection frame B to generate a corresponding first stenosis segment diameter sequence and first stenosis rate;

   Taking the fractional blood flow reserve analysis point marked by the user on the first image as the corresponding target point P _i , and taking the measurement result of the blood vessel diameter at the target point P _i by the user as the corresponding target point diameter; wherein, 1≤i≤n, n is the total number of measurement points;

   Confirm the vascular branch of each target point P _i on the first image to generate a corresponding vascular branch C _i , and perform a vascular branch feature extraction process according to all the first stenotic segment diameter sequences and the first stenosis rate corresponding to each of the vascular branch C _i , the target point diameter of the corresponding target point P _i , and the coronary artery entrance diameter to generate a corresponding branch feature data sequence S _i , and sort all the obtained branch feature data sequences S _i to obtain a branch feature data set D _in (S ₁ ... S _i ... S _n );

   Input the branch feature data set D _in (S ₁ ... S _i ... S _n ) into the preset artificial neural network ANN model for operation and output the branch operation result set D _out (U ₁ ... U _i ... U _n ); and output the operation result U _i of each branch in the set as the analysis result of the blood flow reserve fraction of the corresponding target point P _i .
2. The processing method based on angiographic image analysis blood flow reserve fraction according to claim 1, wherein the image target detection and semantic segmentation model comprises a Mask R-CNN model;

   When the image target detection and semantic segmentation model is specifically the Mask R-CNN model, it includes a feature extraction network layer, a region candidate network layer, a region alignment network layer and a region head network layer;

   The feature extraction network layer is connected to the region candidate network layer, the region candidate network layer is connected to the region alignment network layer, and the region alignment network layer is connected to the region head network layer;

   The feature extraction network layer is specifically composed of a five-level residual network and a corresponding five-level feature pyramid network; the region candidate network layer includes a five-level region candidate network, corresponding to the five-level feature pyramid network; when implementing the five-level residual network, use the residual network ResNet-50 network structure to implement and use this as the backbone network of the feature extraction network layer;

   The regional head network layer includes two sub-networks respectively: a target detection branch network and a target segmentation branch network; the target detection branch network is used to output the target detection frame A whose detection target type is a coronary inlet vessel type, and the target detection frame B whose detection target type is a stenotic segment blood vessel type; the target segmentation branch network is used to output the blood vessel mask image of the corresponding coronary inlet vessel in the target detection frame A, and output the corresponding blood vessel mask image of the narrow segment blood vessel in the target detection frame B.
3. The processing method for analyzing blood flow reserve fraction based on angiographic images according to claim 1, wherein the analysis of the coronary artery entrance diameter on the blood vessel mask image of the target detection frame A to generate a corresponding coronary artery entrance diameter specifically includes:

   Using the blood vessel mask image of the target detection frame A as the current blood vessel mask image;

   Performing blood vessel edge and blood vessel centerline identification on the current blood vessel mask image to generate corresponding current blood vessel edges and current centerline;

   The plurality of pixels included in the current centerline are sequentially marked as corresponding first pixel points X _1,j ; wherein, 1≤j≤m ₁ , m ₁ is the total number of pixels in the current centerline; the first first pixel point X _1,j=1 is the blood flow inflow point of the blood vessel corresponding to the current blood vessel mask image, and the last first pixel point

   

   The blood outflow point of the blood vessel corresponding to the current blood vessel mask image;

   Through each of the first pixel points X _1,j, make a plurality of first intersecting line segments with the edge of the current blood vessel to obtain a plurality of first intersecting line segments; and select the shortest length of the first intersecting line segment as the first pixel point blood vessel diameter L _1,j corresponding to the current first pixel point X _1,X ;

   Perform mean calculation on all obtained first pixel vessel diameters L _1,j to generate the coronary artery entrance diameter.
4. The processing method for analyzing blood flow reserve fraction based on angiographic images according to claim 1, wherein the analysis of the stenotic vessel diameter and stenosis rate on the blood vessel mask image of each target detection frame B generates a corresponding first stenotic segment diameter sequence and first stenosis rate, specifically comprising:

   Using the blood vessel mask image of the current target detection frame B as the current blood vessel mask image;

   Performing blood vessel edge and blood vessel centerline identification on the current blood vessel mask image to generate corresponding current blood vessel edges and current centerline;

   The plurality of pixels included in the current centerline are sequentially marked as corresponding second pixel points X _2,h ; wherein, 1≤h≤m ₂ , m ₂ is the total number of pixels in the current centerline; the first second pixel point X _2,h=1 is the blood flow inflow point of the blood vessel corresponding to the current blood vessel mask image, and the last second pixel point

   

   The blood outflow point of the blood vessel corresponding to the current blood vessel mask image;

   Through each of the second pixel points X _{2, h} do intersection line segments with the current blood vessel edge to obtain a plurality of second intersection line segments; and select the length of the shortest second intersection line segment therefrom as the second pixel point blood vessel diameter L _2,h corresponding to the current second pixel point X _2,h ;

   In addition, according to the second pixel blood vessel diameter L _{2, h=1} and the second pixel blood vessel diameter

   

   The construction can reflect the first second pixel point X _{2, h=1} to the last second pixel point

   

   The linear change function f(h) of the blood vessel linear change relationship,

   f(h)=L _2,h=1 +k*(h-1),

   

   And according to the linear change function f(h), calculate the linear change diameter length corresponding to each second pixel point X _2,h to generate the corresponding second pixel point linear diameter L′ _2,h ,

   

   Then, according to the second pixel blood vessel diameter L _2,h and the second pixel linear diameter L′ _2,h , calculate the second pixel stenosis rate R 2 _,h corresponding to each second pixel X _2,h ,

   

   Using the second pixel point blood vessel diameter L _2,h as the corresponding first stenosis segment diameter, sorting all the first stenosis segment diameters in ascending order according to the second pixel point index h, generating the first stenosis segment diameter sequence corresponding to the current blood vessel mask image; and selecting the maximum value from all the obtained second pixel point stenosis rates R _2,h as the first stenosis rate corresponding to the current blood vessel mask image.
5. The method for processing blood flow reserve fraction based on angiographic image analysis according to claim 1, wherein the blood vessel branch of each target point P _i on the first image is confirmed to generate a corresponding blood vessel branch C _i , and according to all the first stenotic segment diameter sequences and the first stenosis rate corresponding to each of the blood vessel branches C _i , and the corresponding target point diameter of the target point _Pi and the coronary artery entrance diameter, a blood vessel branch feature extraction process is performed to generate a corresponding branch feature data sequence S _i , and all the obtained branch feature data are obtained. The sequence S _i is sorted to obtain the branch feature data set D _in (S ₁ ... S _i ... S _n ), which specifically includes:

   On the first image, the starting position of the blood vessel in the blood vessel mask image of the target detection frame A is used as the coronary artery inlet position; and according to the blood flow direction, the blood flow path from the coronary artery inlet position to each of the target points P _i is used as the blood vessel branch C _i corresponding to each of the target points P _i ;

   Classifying the first stenotic segment diameter sequences of all the target detection frames B passed by each of the blood vessel branches _Ci into corresponding first sequence sets; and in each of the first sequence sets, sorting the first stenotic segment diameters of all the first stenotic segment diameter sequences according to the blood flow direction to obtain the corresponding first branch stenotic segment diameter sequences;

   From the first stenosis rates of all the target detection frames B passed by each of the blood vessel branches _Ci , select a maximum value as the corresponding maximum stenosis rate of the first branch;

   Taking the target diameter of the target point P _i corresponding to each of the blood vessel branches C _i as the corresponding first branch target diameter;

   For each of the blood vessel branches C _i , the coronary inlet diameter and the corresponding first branch stenosis segment diameter sequence, the first branch target point diameter and the first branch maximum stenosis rate are combined to form the corresponding branch characteristic data sequence S _i ;

   All the branch feature data sequences S _i are sorted in ascending order of the target point index i, so as to obtain the branch feature data set D _in (S ₁ ... S _i ... S _n ).
6. The processing method for analyzing blood flow reserve fractions based on angiographic images according to claim 1, wherein the branch feature data set D _in (S ₁ ... S _i ... S _n ) is input into a preset artificial neural network ANN model for calculation and output branch operation result set D _out (U ₁ ... U _i ... U _n ), specifically comprising:

   Dividing the branch feature data set D _in (S ₁ ... S _i ... S _n ) into n one-dimensional first data segments; each first data segment corresponds to a branch feature data sequence S _i ;

   Input each of the first data fragments into the artificial neural network (ANN) model for calculation to obtain the corresponding branch calculation result U _i ;

   All the branch operation results U _i are sorted in ascending order of the target point index i, so as to obtain the branch operation result set D _out (U ₁ ... U _i ... U _n ).
7. The processing method for analyzing blood flow reserve fraction based on angiographic images according to claim 6, characterized in that,

   The artificial neural network ANN model includes an input layer, one or more hidden layers and an output layer; the input layer includes a plurality of input layer nodes; each of the hidden layers includes a plurality of hidden layer nodes; the output layer includes an output layer node;

   The input layer is used to input each segment data of the first data segment to the corresponding input layer node as a corresponding node output value;

   Each of the hidden layer nodes of the first hidden layer is fully connected with all the input layer nodes to form a corresponding first fully connected network; the first hidden layer is used to input the node output values of all the input layer nodes in the corresponding first fully connected network into a preset fully connected linear operation function to generate a corresponding first result, and input the first result into a preset activation function to generate a corresponding second result, and use the second result as the node output value of the current hidden layer node on each of the hidden layer nodes of this layer;

   Each of the hidden layer nodes of the next hidden layer is fully connected with all the hidden layer nodes of the previous hidden layer to form a corresponding second fully connected network; the next hidden layer is used to input the node output values of all the hidden layer nodes of the upper hidden layer in the corresponding second fully connected network into a preset fully connected linear operation function to generate a corresponding third result, and input the third result into a preset activation function to perform calculation to generate a corresponding fourth result, and use the fourth result as the node output value of the current hidden layer node. ;

   The output layer nodes of the output layer are fully connected with all the hidden layer nodes of the last hidden layer to form a corresponding third fully connected network; the output layer is used to input, on the output layer nodes, the node output values of all the hidden layer nodes of the last hidden layer in the corresponding third fully connected network into a preset fully connected linear operation function for operation to generate a corresponding fifth result, and input the fifth result into a preset activation function for operation to generate a corresponding sixth result, and use the sixth result as the branch operation result U _i ;

   The activation function used by the hidden layer and the output layer is a ReLU function by default.
8. A device for implementing the steps of the processing method for analyzing blood flow reserve based on angiographic image analysis according to any one of claims 1-7, wherein the device comprises: an acquisition module, an image target detection and semantic segmentation module, a quantitative coronary angiography analysis and processing module, a target point processing module, a blood vessel branch processing module, and a blood flow reserve processing module;

   The obtaining module is used to obtain an angiographic image as a first image;

   The image target detection and semantic segmentation module is used to perform target detection and semantic segmentation processing of coronary inlet vessels and stenotic vessels on the first image based on a preset image target detection and semantic segmentation model, so as to obtain a target detection frame A and one or more target detection frames B on the first image; each of the target detection frames A and B includes a section of blood vessel mask image, the detection target type corresponding to the target detection frame A is the coronary artery entrance blood vessel type, and the detection target type corresponding to the target detection frame B is the stenotic segment blood vessel type;

   The quantitative coronary angiography analysis and processing module is used to analyze the coronary artery entrance diameter on the blood vessel mask image of the target detection frame A to generate a corresponding coronary artery entrance diameter; and analyze the stenotic segment blood vessel diameter and stenosis rate analysis on the blood vessel mask image of each target detection frame B to generate a corresponding first stenosis segment diameter sequence and first stenosis rate;

   The target point processing module is used to use the blood flow reserve fraction analysis point marked by the user on the first image as the corresponding target point P _i , and use the user's blood vessel diameter measurement result at the target point P _i as the corresponding target point diameter; where 1≤i≤n, n is the total number of measurement points;

   The blood vessel branch processing module is used to confirm the blood vessel branch of each target point P _i on the first image to generate a corresponding blood vessel branch C _i , and perform blood vessel branch feature extraction processing according to all the first stenosis segment diameter sequences and the first stenosis rate corresponding to each of the blood vessel branch C _i , the target point diameter of the corresponding target point _Pi , and the coronary artery entrance diameter to generate a corresponding branch feature data sequence S _i , and sort all the obtained branch feature data sequences S _i to obtain a branch feature data set D _in (S ₁ ... S _i ... S _n );

   The blood flow reserve fraction processing module is used to input the branch feature data set D _in (S ₁ ... S _i ... S _n ) into the preset artificial neural network ANN model to perform calculation and output a branch operation result set D _out (U ₁ ... U _i ... U _n ); and output each branch operation result U _i in the set as the corresponding blood flow reserve fraction analysis result of the target point P _i .
9. An electronic device, characterized in that it includes: a memory, a processor, and a transceiver;

   The processor is configured to be coupled to the memory, read and execute instructions in the memory, so as to implement the method steps described in any one of claims 1-7;

   The transceiver is coupled to the processor, and the processor controls the transceiver to send and receive messages.
10. A computer-readable storage medium, characterized in that the computer-readable storage medium stores computer instructions, and when the computer instructions are executed by a computer, the computer executes the instructions of the method according to any one of claims 1-7.

Images