CN113744246B

CN113744246B - Method and apparatus for predicting fractional flow reserve from a vessel tomographic image

Info

Publication number: CN113744246B
Application number: CN202111034145.XA
Authority: CN
Inventors: 夏天祎; 曹君
Original assignee: Lepu Medical Technology Beijing Co Ltd
Current assignee: Lepu Medical Technology Beijing Co Ltd
Priority date: 2021-09-03
Filing date: 2021-09-03
Publication date: 2023-07-18
Anticipated expiration: 2041-09-03
Also published as: CN113744246A

Abstract

Embodiments of the present invention relate to a method and apparatus for predicting fractional flow reserve from a tomographic image of a blood vessel, the method comprising: acquiring a first cross-sectional sequence of first tomographic images; performing cross-sectional area calculation and blood movement distance calculation on each first cross section to generate a first area-distance data set; performing curve conversion on the first area-distance data set sequence to generate a first curve; carrying out vascular stenosis identification on the first curve to generate a plurality of first stenosis; recording the curve segments separated by the respective first narrow segments as first non-narrow segments; performing a first fractional flow reserve variation prediction to generate first variation data; performing a second fractional flow reserve variation prediction to generate second variation data; sequencing the first variable data and the second variable data to generate a variable data sequence; a fractional flow reserve prediction process is performed to generate a first fractional flow reserve data sequence. The invention can reduce the detection difficulty and improve the detection safety.

Description

Method and apparatus for predicting fractional flow reserve from a vessel tomographic image

Technical Field

The invention relates to the technical field of data processing, in particular to a method and a device for predicting fractional flow reserve according to a blood vessel tomography image.

Background

Fractional flow reserve (Fractional Flow Reserve, FFR) refers to the ratio of the average pressure within the stenosed distal coronary artery to the average pressure of the coronary ostial aorta at maximum hyperemia of the coronary vessel. FFR values are conventionally obtained by means of percutaneous coronary intervention (Percutaneous Coronary Intervention, PCI). However, PCI detection is not only complicated to operate but also an invasive detection means, and may cause a certain physical damage to the detection subject, with a certain risk.

Disclosure of Invention

The invention aims at overcoming the defects of the prior art, and provides a method, a device, electronic equipment and a computer readable storage medium for predicting fractional flow reserve according to a blood vessel tomography image, wherein a blood vessel cross section sampling sequence of a three-dimensional blood vessel tomography image obtained by using a computer tomography angiography (Computed Tomography Angiography, CTA) technology is subjected to estimation of a blood vessel cross section area and a blood flow movement distance, a curve reflecting an area-distance corresponding relation is generated according to the estimation result, a blood vessel narrow section and a non-narrow section are identified according to the area-distance curve, FFR variation of each narrow section and each non-narrow section is predicted, and the FFR value of each blood vessel cross section of a blood vessel is predicted according to the predicted FFR variation. By the method, personal injury caused by invasive examination can be avoided, the detection difficulty can be greatly reduced, and the detection safety is improved.

To achieve the above object, a first aspect of an embodiment of the present invention provides a method of predicting fractional flow reserve from a blood vessel tomographic image, the method comprising:

acquiring a first cross-sectional sequence of first tomographic images; the first tomographic image is a three-dimensional tomographic image; the first sequence of cross sections includes a plurality of first cross sections; the first section is sampling information obtained by sampling a blood vessel cross section of the first tomographic image along a blood flow direction and being perpendicular to the blood flow direction;

calculating the cross-sectional area of each first cross section to generate corresponding first area data, calculating the blood movement distance to generate corresponding first distance data, and forming a corresponding first area-distance data set by the obtained first area data and the first distance data; and forming a first sequence of area-distance data sets from the resulting plurality of sets of first area-distance data sets;

performing curve conversion processing on the first area-distance data set sequence by taking the area as an ordinate and the distance as an abscissa to generate a first curve;

carrying out recognition processing on the vascular stenosis sections on the first curve to generate a plurality of first stenosis sections; and recording the curve segments of said first curve separated by each of said first narrow segments as first non-narrow segments;

Performing first fractional flow reserve variation prediction processing on each first narrow section to generate corresponding first variation data; performing second fractional flow reserve variation prediction processing on each first non-stenosis segment to generate corresponding second variation data;

sequencing the obtained plurality of first variable quantity data and the second variable quantity data according to the sequence of each first narrow section and each first non-narrow section to generate a variable quantity data sequence;

and carrying out fractional flow reserve prediction processing on the section fractional flow reserve according to the variable quantity data sequence to generate a first fractional flow reserve data sequence.

Preferably, the first section includes a plurality of first section edge point coordinates; the calculating the cross-sectional area of each first cross-section to generate corresponding first area data, calculating the blood movement distance to generate corresponding first distance data, and forming a corresponding first area-distance data set by the obtained first area data and the first distance data specifically includes:

performing closed curve fitting processing according to a plurality of first section edge point coordinates of a current first section to generate a corresponding first closed curve graph;

Performing graph area estimation processing on the first closed curve graph to generate first area data;

performing graph center point estimation processing on the first closed curve graph to generate a first section center point coordinate;

calculating the blood movement distance from the first section center point coordinate to the current first section center point coordinate along the blood flow direction, and generating the first distance data;

the first area-distance data set is configured from the first area data and the first distance data.

Preferably, the identifying the first curve to a stenosis of a blood vessel, generating a plurality of first stenosis specifically includes:

identifying preset treatment mode data of the narrow section of the blood vessel;

if the blood vessel stenosis processing mode data is a first mode, identifying peak points of the first curve to generate a plurality of first peak points; and the curves between the adjacent first peak points are marked as first curve interval sections; absolute difference value calculation is carried out on the distance coordinate values of the first coordinate point and the last coordinate point of the first curve interval section to generate first interval section distance data; average value calculation is carried out on the area coordinate values of the first coordinate point and the last coordinate point of the first curve interval section to generate first average area data, the minimum area coordinate value of the first curve interval section is extracted to generate first minimum area data, and the ratio of the first minimum area data to the first average area data is calculated to generate a first area ratio; when the first interval distance data is higher than a preset interval distance threshold value and the first area ratio is lower than a preset area ratio characteristic threshold value, marking the first curve interval section as the first narrow section;

If the data of the treatment mode of the vascular stenosis section is the second mode, drawing a two-dimensional vascular shape graph according to the first curve to generate a first vascular shape graph; inputting the first vessel shape graph into a first intelligent model which is mature in training for first vessel stenosis segmentation treatment, and obtaining a first mark shape graph with a plurality of vessel stenosis region marks; and in the first curve, marking a curve interval section corresponding to each blood vessel narrow section area mark of the first mark shape chart as the first narrow section;

if the blood vessel stenosis processing mode data is a third mode, performing second blood vessel stenosis segmentation processing on the first tomographic image corresponding to the first curve by using a second intelligent model which is mature in training, so as to obtain a first marked tomographic image with a plurality of blood vessel stenosis region marks; and in the first curve, a curve interval section corresponding to each blood vessel stenosis section area mark of the first mark tomographic image is recorded as the first stenosis section.

Further, extracting a first area-distance training curve from a preset training dataset before using the first smart model; drawing a two-dimensional blood vessel shape graph according to the first area-distance training curve to generate a first training blood vessel shape graph; performing artificial blood vessel stenosis region marking processing on the first training blood vessel shape graph to generate a first training mark shape graph with a plurality of blood vessel stenosis region marks; inputting the first training blood vessel shape graph into the first intelligent model for first blood vessel stenosis segmentation processing, and generating a second training blood vessel shape graph with a plurality of blood vessel stenosis region marks; performing error calculation on the blood vessel stenosis region marking results of the first training marking shape graph and the second training blood vessel shape graph, and performing reverse modulation on the first intelligent model according to the error calculation result until the error calculation result enters a preset first model error convergence range;

Extracting a second area-distance training curve from a preset training dataset before using the second smart model; extracting a three-dimensional tomographic image corresponding to the second area-distance training curve from the training data set, and recording the three-dimensional tomographic image as a first training tomographic image; performing artificial fractional flow reserve calculation on the first training tomographic image according to a computational fluid dynamics method to generate a first training fractional flow reserve data sequence; performing artificial blood vessel stenosis region marking processing on the first training tomographic image according to the first training blood flow reserve fraction data sequence to generate a first training marking tomographic image with a plurality of blood vessel stenosis region marks; inputting the first training tomographic image into the second intelligent model for second blood vessel stenosis segmentation processing, and generating a second training mark tomographic image with a plurality of blood vessel stenosis region marks; and performing error calculation on the blood vessel stenosis region marking results of the first training mark tomographic image and the second training mark tomographic image, and performing reverse modulation on the second intelligent model according to the error calculation result until the error calculation result enters a preset second model error convergence range.

Preferably, the first fractional flow reserve variation prediction processing is performed on each first stenosis, and the generating of the corresponding first variation data specifically includes:

analyzing the shape characteristics of the current first narrow section to generate corresponding first narrow section shape characteristic data; analyzing the narrow section position characteristics of the current first narrow section to generate corresponding first narrow section position characteristic data; analyzing the shape characteristics of the blood vessel where the current first stenosis is located, and generating corresponding first blood vessel shape characteristic data;

inputting the first stenosis shape characteristic data, the first stenosis position characteristic data and the first vessel shape characteristic data into a training mature stenosis fractional flow reserve variation prediction model to perform stenosis fractional flow reserve variation prediction processing, and generating corresponding first variation data.

Preferably, the second fractional flow reserve variation prediction processing is performed on each of the first non-stenosis sections to generate corresponding second variation data, and specifically includes:

analyzing the shape characteristics of the current first non-narrow section to generate corresponding first non-narrow section shape characteristic data; analyzing the shape characteristics of the blood vessel where the current first non-stenosis section is located, and generating corresponding second blood vessel shape characteristic data;

And inputting the first non-stenosis segment shape characteristic data and the second blood vessel shape characteristic data into a training mature non-stenosis segment fractional flow reserve variation prediction model to perform non-stenosis segment fractional flow reserve variation prediction processing, and generating the corresponding second variation data.

Preferably, the variable data sequence includes a plurality of third variable data; the step of performing fractional flow reserve prediction processing on the cross section fractional flow reserve according to the variable data sequence to generate a first fractional flow reserve data sequence specifically comprises the following steps:

recording a narrow section or a non-narrow section of the first curve corresponding to the first third variable quantity data as a first curve section; classifying a plurality of first sections corresponding to the first curve segments into a first section set; and storing first fractional flow reserve data FFR corresponding to a first one of the first cross-sections of the first set of cross-sections _1F Setting a preset initial Fractional Flow Reserve (FFR) and obtaining first fractional flow reserve data FFR corresponding to the last first section of the first section set _1L Set as (FFR) _1F +first third variation data); recording said first cross-section except for the first and last one of said first set of cross-sections as a first intermediate cross-section; setting the fractional flow reserve data of the first section corresponding to any one of the first middle sections as a1 is the distance from the first intermediate section to the first one of the first set of sections, b1 is the distance from the first intermediate section to the last one of the first set of sections, c1 is the distance from the first one of the first set of sections to the last one of the first section;

recording a narrow section or a non-narrow section of the first curve corresponding to the second variable quantity data as a second curve section; classifying a plurality of first sections corresponding to the second curve segments into a second section set; and storing first fractional flow reserve data FFR corresponding to a first one of the first cross-sections of the second set of cross-sections _2F First cross-sectional fractional flow reserve data FFR corresponding to the last of the first cross-sections set of the first cross-sections _1L First cross-sectional fractional flow reserve data FFR corresponding to the last of the first cross-sections of the second set of cross-sections _2L Set as (FFR) _2F +second third variation data); dividing the second set of cross sections by the first oneAnd the first section other than the last is marked as a second intermediate section; setting the fractional flow reserve data of the first section corresponding to any one of the second intermediate sections as a2 is the distance from the second intermediate section to the first one of the second set of sections, b2 is the distance from the second intermediate section to the last one of the second set of sections, c2 is the distance from the first one of the second set of sections to the last one of the first sections; and so on until the last of the third variation data;

and sequencing all the obtained first section fractional flow reserve data according to the sequence, and generating the first section fractional flow reserve data sequence.

A second aspect of an embodiment of the present invention provides an apparatus for implementing the method described in the first aspect, including: the device comprises an acquisition module, a data processing module, a stenosis and non-stenosis identification module and a fractional flow reserve prediction module;

the acquisition module is used for acquiring a first section sequence of a first tomographic image; the first tomographic image is a three-dimensional tomographic image; the first sequence of cross sections includes a plurality of first cross sections; the first section is sampling information obtained by sampling a blood vessel cross section of the first tomographic image along a blood flow direction and being perpendicular to the blood flow direction;

The data processing module is used for calculating the cross-sectional area of each first cross section to generate corresponding first area data, calculating the blood movement distance to generate corresponding first distance data, and forming a corresponding first area-distance data set by the obtained first area data and the first distance data; and forming a first sequence of area-distance data sets from the resulting plurality of sets of first area-distance data sets; and performing curve conversion processing on the first area-distance data set sequence by taking the area as an ordinate and the distance as an abscissa to generate a first curve;

the stenosis and non-stenosis identification module is used for carrying out vessel stenosis identification processing on the first curve to generate a plurality of first stenosis; and recording the curve segments of said first curve separated by each of said first narrow segments as first non-narrow segments;

the fractional flow reserve prediction module is used for performing first fractional flow reserve variation prediction processing on each first narrow section to generate corresponding first variation data; performing second fractional flow reserve variation prediction processing on each first non-stenosis segment to generate corresponding second variation data; sequencing the obtained plurality of first variable quantity data and the second variable quantity data according to the sequence of each first narrow section and each first non-narrow section to generate a variable quantity data sequence; and performing fractional flow reserve prediction processing on the cross-section fractional flow reserve according to the variable data sequence to generate a first fractional flow reserve data sequence.

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

the processor is configured to couple to the memory, and read and execute the instructions in the memory, so as to implement the method steps described in the first aspect;

the transceiver is coupled to the processor and is controlled by the processor to transmit and receive messages.

A fourth aspect of the embodiments of the present invention provides a computer-readable storage medium storing computer instructions that, when executed by a computer, cause the computer to perform the instructions of the method of the first aspect.

The embodiment of the invention provides a method, a device, electronic equipment and a computer readable storage medium for predicting fractional flow reserve according to a blood vessel tomographic image, which are used for estimating the blood vessel cross section area and the blood flow movement distance of a blood vessel cross section sampling sequence of a three-dimensional blood vessel tomographic image obtained by using a CTA technology, generating a curve reflecting the corresponding relation of the area and the distance according to the estimation result, identifying a narrow section and a non-narrow section of the blood vessel according to the area-distance curve, predicting FFR variation of each narrow section and each non-narrow section, and finally predicting the FFR value of each blood vessel cross section according to each predicted FFR variation. By the method, personal injury caused by invasive examination is avoided, the detection difficulty is greatly reduced, and the detection safety is improved.

Drawings

FIG. 1 shows a blood vessel according to a first embodiment of the present invention a schematic diagram of a method for predicting fractional flow reserve by using tomographic images;

FIG. 2 is a block diagram of an apparatus for predicting fractional flow reserve based on a tomographic image of a blood vessel according to a second embodiment of the present invention;

fig. 3 is a schematic structural diagram of an electronic device according to a third embodiment of the present invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in further detail below with reference to the accompanying drawings, and it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.

As shown in fig. 1, which is a schematic diagram of a method for predicting fractional flow reserve according to a blood vessel tomography image according to a first embodiment of the present invention, the method mainly includes the following steps:

step 1, acquiring a first section sequence of a first tomographic image;

Wherein the first tomographic image is a three-dimensional tomographic image; the first cross-section sequence includes a plurality of first cross-sections; the first section is sampling information obtained by sampling a blood vessel cross section of the first tomographic image along the blood flow direction, wherein the blood vessel cross section is perpendicular to the blood flow direction; the first section includes a plurality of first section edge point coordinates.

Here, the first tomographic image is a three-dimensional tomographic angiographic image obtained using a CTA technique; when the first tomographic image is sampled along the blood flow direction on the blood vessel cross section perpendicular to the blood flow direction, the coordinates of each point on the blood vessel wall, namely the edge point on the blood vessel cross section, are taken as corresponding first cross section edge point coordinates, and then the first cross section is an information sequence consisting of a plurality of first cross section edge point coordinates; the direction of blood flow here is the direction of blood movement from the proximal end of the vessel to the distal end of the vessel.

Step 2, calculating the cross-sectional area of each first cross section to generate corresponding first area data, calculating the blood movement distance to generate corresponding first distance data, and forming a corresponding first area-distance data set by the obtained first area data and the first distance data; and forming a first area-distance data set sequence from the plurality of obtained first area-distance data sets;

Here, the first area data is area information of each blood vessel cross section, that is, a first cross section, the first distance data is a distance between each first cross section and the blood vessel inlet, that is, a 1 st first cross section, the distance is not a straight line distance but a movement distance of blood in the blood vessel from the blood vessel inlet to a current first cross section, the first area-distance data set can represent a corresponding relationship between the blood vessel cross section and a blood movement distance on the corresponding first cross section, and the first area-distance data set sequence can represent a corresponding relationship between the blood vessel cross sections and the blood movement distances on different positions on the whole blood vessel;

the method specifically comprises the following steps: step 21, performing closed curve fitting processing according to a plurality of first section edge point coordinates of a current first section to generate a corresponding first closed curve graph;

here, the first closed curve pattern is an irregular pattern;

step 22, performing graph area estimation processing on the first closed curve graph to generate first area data;

here, the graph area estimation is performed based on an integration method of the first closed curve graph edge point and the center point so as to obtain first area data;

step 23, performing graph center point estimation processing on the first closed curve graph to generate a first section center point coordinate;

Here, the coordinates of all the edge points on each coordinate axis are averaged to be used as the coordinates of the center point of the first section;

step 24, calculating the blood movement distance from the first section center point coordinate to the current first section center point coordinate along the blood flow direction, and generating first distance data;

here, the first distance data corresponding to the first section should be 0;

when calculating the blood movement distance from the first cross section center point coordinate to the second first cross section center point coordinate, firstly calculating the linear distance between the second first cross section center point coordinate and the first cross section center point coordinate to generate first relative distance data, and then adding the first distance data corresponding to the previous first cross section (the first distance data corresponding to the first cross section) and the first relative distance data to obtain first distance data corresponding to the second first cross section;

when calculating the blood movement distance from the first section center point coordinate to the third first section center point coordinate, firstly calculating the linear distance between the third first section center point coordinate and the second first section center point coordinate to generate second relative distance data, and then adding the first distance data corresponding to the previous first section (the first distance data corresponding to the second first section) and the second relative distance data to be used as the first distance data corresponding to the third first section;

And the like, until first distance data corresponding to the last first section are obtained;

step 25, forming a corresponding first area-distance data set by the first area data and the first distance data;

here, the first area-distance data set reflects a correspondence between the blood vessel cross section and the blood movement distance on the corresponding first section;

step 26, forming a first sequence of area-distance data sets from the resulting plurality of sets of first area-distance data sets.

Here, the first area-distance data set sequence represents a correspondence of a blood vessel cross section at different locations over the whole blood vessel with a blood movement distance.

And 3, performing curve conversion processing on the first area-distance data set sequence by taking the area as an ordinate and the distance as an abscissa to generate a first curve.

When curve conversion processing is carried out on the first area-distance data set sequence, the first area data of each first area-distance data set in the first area-distance data set sequence is taken as an ordinate value, the first distance data is taken as an abscissa value, an area-distance coordinate point mark is carried out on a two-dimensional coordinate system taking the area as an ordinate and the distance as an abscissa to obtain a plurality of area-distance coordinate points, and the area-distance coordinate points are sequentially connected to obtain an initial curve;

In order to reduce the noise of the initial curve, a filter processing method aiming at the aperiodic signal can be further adopted to carry out curve smoothing processing on the initial curve, so as to obtain a first curve capable of reflecting the change trend of the blood vessel cross section and the blood movement distance.

Step 4, carrying out recognition processing on the vascular stenosis sections on the first curve to generate a plurality of first stenosis sections; and recording the curve segments of the first curve separated by the respective first narrow segments as first non-narrow segments;

step 41, carrying out vascular stenosis identification processing on the first curve to generate a plurality of first stenosis;

here, the embodiment of the invention provides three ways of identifying the stenosis of the blood vessel, which correspond to three modes of preset system parameter of the stenosis processing mode data of the blood vessel respectively: a first mode, a second mode, and a third mode;

the method specifically comprises the following steps: step 411, identifying preset vascular stenosis treatment mode data; if the stenotic lesion treatment pattern data is the first pattern, then go to step 412; if the data of the treatment mode of the blood vessel stenosis is the second mode, go to step 413; if the stenosis treatment mode data is the third mode, go to step 414;

here, if the data of the treatment mode of the blood vessel stenosis is the first mode, the first recognition mode of the blood vessel stenosis is adopted, and the following step 412 is performed; if the data of the treatment mode of the blood vessel stenosis is the second mode, the second recognition mode of the blood vessel stenosis is adopted currently, and the following step 413 is used for treatment; if the data of the treatment mode of the blood vessel stenosis is the third mode, the third recognition mode of the blood vessel stenosis is adopted currently, and the following step 414 is used for treatment;

Step 412, performing identification processing on the peak points of the first curve to generate a plurality of first peak points; and the curves between the adjacent first peak points are marked as first curve interval sections; absolute difference value calculation is carried out on the distance coordinate values of the first coordinate point and the last coordinate point of the first curve interval section to generate first interval section distance data; average value calculation is carried out on the area coordinate values of the first coordinate point and the last coordinate point of the first curve interval section to generate first average area data, the minimum area coordinate value of the first curve interval section is extracted to generate first minimum area data, and the ratio of the first minimum area data to the first average area data is calculated to generate a first area ratio; when the first interval distance data is higher than a preset interval distance threshold value and the first area ratio is lower than a preset area ratio characteristic threshold value, marking the first curve interval section as a first narrow section; go to step 42;

when the first vascular stenosis recognition method is used for processing, firstly, taking peak points of a first curve as sub-segment dividing conditions, taking a region between every two peak points as a sub-segment, namely a first curve interval segment, wherein the head and the tail of each first curve interval segment respectively correspond to an area-distance coordinate point;

And then carrying out vascular stenosis judgment on each first curve interval section:

firstly, calculating absolute difference values of coordinate values of a transverse axis, namely distance coordinate values of the first coordinate point and the second coordinate point to obtain first interval distance data;

for example, the head-to-tail coordinate points of the current first curve interval segment are respectively: coordinate point 1 (distance 1, area 1) and coordinate point 2 (distance 2, area 2), the first interval distance data= |distance 2-distance 1|, wherein||is absolute value sign;

secondly, calculating the average value of the vertical axis coordinate values, namely the area coordinate values, of the first coordinate point and the second coordinate point to obtain first average area data; extracting a minimum vertical axis coordinate value, namely a minimum area coordinate value, from the first curve interval section as first minimum area data; and calculating a first area ratio = first minimum area data/first average area data from the first minimum area data and the first average area data;

for example, the head-to-tail coordinate points of the current first curve interval segment are respectively: coordinate point 1 (distance 1, area 1) and coordinate point 2 (distance 2, area 2), the minimum area coordinate value of the current first curve interval corresponds to coordinate point 3 (distance 3, area 3), then first average area data= (area 1+area 2)/2, first minimum area data = area 3, first area ratio = first minimum area data/first average area data = area 3/((area 1+area 2)/2) =2×area 3/(area 1+area 2);

Finally, judging the first interval distance data and the first area ratio according to preset interval distance threshold values, area ratio characteristic threshold values and other characteristic threshold parameters, and if the first interval distance data is more than or equal to the interval distance threshold value and the first area ratio is less than or equal to the area ratio characteristic threshold value, determining the current first curve interval as a blood vessel stenosis section and marking the current first curve interval as a first stenosis section;

step 413, drawing a two-dimensional vascular shape graph according to the first curve to generate a first vascular shape graph; inputting the first vessel shape graph into a first intelligent model which is mature in training for carrying out first vessel stenosis segmentation treatment to obtain a first mark shape graph with a plurality of vessel stenosis region marks; in the first curve, curve interval sections corresponding to the area marks of the narrow sections of the blood vessels in the first mark shape chart are marked as first narrow sections; go to step 42;

further, extracting a first area-distance training curve from a preset training dataset before using the first smart model; drawing a two-dimensional blood vessel shape graph according to the first area-distance training curve to generate a first training blood vessel shape graph; carrying out artificial blood vessel stenosis region marking treatment on the first training blood vessel shape graph to generate a first training mark shape graph with a plurality of blood vessel stenosis region marks; inputting the first training blood vessel shape graph into a first intelligent model for first blood vessel stenosis segmentation processing, and generating a second training blood vessel shape graph with a plurality of blood vessel stenosis region marks; performing error calculation on the blood vessel stenosis region marking results of the first training marking shape graph and the second training blood vessel shape graph, and performing reverse modulation on the first intelligent model according to the error calculation result until the error calculation result enters a preset first model error convergence range;

When the second blood vessel stenosis recognition method is used for processing, a two-dimensional blood vessel shape graph, namely a first blood vessel shape graph, is constructed by referring to the corresponding relation between the blood vessel cross-sectional area and the blood movement distance reflected by the first curve, and the difference between the graph and the first tomographic image is that the first blood vessel shape graph has no background noise, and the image recognition result is more accurate; then inputting the first vessel shape graph into a training mature two-dimensional image vessel stenosis semantic recognition model, namely a first intelligent model, and performing two-dimensional image vessel stenosis semantic segmentation processing, namely first vessel stenosis segmentation processing, so as to obtain a semantic recognition image with a plurality of vessel stenosis object mark frames, namely vessel stenosis area marks, namely a first mark shape graph; then, marking a curve interval section corresponding to each blood vessel narrow section area mark in the first curve as a first narrow section;

the first intelligent model is actually the model type capable of realizing the end-to-end two-dimensional image semantic segmentation function in the artificial intelligent model, and can be selected and configured according to the calculation resources, calculation efficiency and other factors of the model in the specific implementation; before using the first intelligent model, the model needs to be subjected to supervision training; during training, a first training mark shape diagram of the artificial blood vessel stenosis region mark is adopted to monitor model output, namely a second training blood vessel shape diagram;

Step 414, performing a second vessel stenosis segmentation process on the first tomographic image corresponding to the first curve by using a second intelligent model which is mature in training, so as to obtain a first marked tomographic image with a plurality of vessel stenosis region marks; in the first curve, curve interval sections corresponding to the region marks of the narrow sections of the blood vessels of the first mark tomographic image are marked as first narrow sections;

further, extracting a second area-distance training curve from a preset training dataset before using the second smart model; extracting a three-dimensional tomographic image corresponding to the second area-distance training curve from the training data set, and recording the three-dimensional tomographic image as a first training tomographic image; performing artificial fractional flow reserve calculation on the first training tomographic image according to a computational fluid dynamics method to generate a first training fractional flow reserve data sequence; performing artificial blood vessel stenosis region marking processing on the first training tomographic image according to the first training blood flow reserve fraction data sequence to generate a first training marking tomographic image with a plurality of blood vessel stenosis region marks; inputting the first training tomographic image into a second intelligent model for second blood vessel stenosis segmentation processing, and generating a second training mark tomographic image with a plurality of blood vessel stenosis region marks; performing error calculation on the blood vessel stenosis region marking results of the first training mark tomographic image and the second training mark tomographic image, and performing reverse modulation on the second intelligent model according to the error calculation result until the error calculation result enters a preset second model error convergence range;

When the third vascular stenosis segment recognition mode is used for processing, inputting a three-dimensional tomographic image corresponding to the first curve, namely a first tomographic image, into a three-dimensional image vascular stenosis semantic recognition model based on a computational fluid dynamics method which is trained and mature, namely a second intelligent model, and performing three-dimensional image vascular stenosis semantic segmentation processing, namely second vascular stenosis segment segmentation processing, so as to obtain a semantic recognition image with a plurality of vascular stenosis segment object marking frames, namely vascular stenosis segment area marks, namely a first marking tomographic image; then, marking a curve interval section corresponding to each blood vessel narrow section area mark in the first curve as a first narrow section;

the second intelligent model is actually the model type capable of realizing the end-to-end three-dimensional image semantic segmentation function in the artificial intelligent model, and can be selected and configured according to the calculation resources, calculation efficiency and other factors of the model in the specific implementation; before using the second intelligent model, the model needs to be supervised and trained; during training, manually calculating the fractional flow reserve data corresponding to each point on the blood vessel in the first training tomographic image by using a related calculation tool according to a computational fluid dynamics method to obtain a corresponding first training fractional flow reserve data sequence, and performing artificial blood vessel stenosis region marking processing on the first training tomographic image according to the first training fractional flow reserve data sequence to obtain supervision data with a plurality of blood vessel stenosis region marks, namely a first training mark tomographic image; monitoring the model output, namely the second training mark tomographic image, by using the first training mark tomographic image;

The curve segments of the first curve separated by respective first narrow segments are denoted as first non-narrow segments, step 42.

For example, on the first curve, a curve interval segment between the coordinate points a to B is identified as a first narrow segment, and then a curve segment between the 1 st coordinate point to the coordinate point a and a curve segment between the coordinate point B to the last 1 coordinate points on the first curve are both identified as a first non-narrow segment.

Step 5, carrying out first fractional flow reserve variation prediction processing on each first narrow section to generate corresponding first variation data; performing second fractional flow reserve variation prediction processing on each first non-stenosis segment to generate corresponding second variation data;

the first variable quantity data refers to the absolute difference value of the fractional flow reserve at the outlet and the fractional flow reserve at the inlet of the blood vessel corresponding to the first narrow section; the second variation data refers to the absolute difference between the fractional flow reserve at the outlet and the fractional flow reserve at the inlet of the blood vessel corresponding to the first non-stenosis segment;

step 51, performing first fractional flow reserve variation prediction processing on each first stenosis segment to generate corresponding first variation data;

here, when the first fractional flow reserve variation prediction processing is performed on each first stenosis, the embodiment of the present invention predicts the fractional flow reserve variation of the current first stenosis according to the stenosis shape feature, the stenosis position feature and the vessel shape feature of the vessel in which each first stenosis is located;

The method specifically comprises the following steps: step 511, analyzing the shape characteristics of the current first stenosis to generate corresponding first stenosis shape characteristic data; analyzing the position characteristics of the current first narrow section to generate corresponding first narrow section position characteristic data; analyzing the shape characteristics of the blood vessel where the current first stenosis is located, and generating corresponding first blood vessel shape characteristic data;

the first stenosis shape characteristic data is a shape information set of each first stenosis, the first stenosis position characteristic data is an information set of the relative position of each first stenosis on the vessel tree, and the first vessel shape characteristic data is an information set of the relative position of each first stenosis on the vessel tree;

the three analysis processes are specifically as follows:

(1) When analyzing the stenosis shape characteristics of the current first stenosis:

step A-1, marking an entrance section, a narrowest section and an exit section of a current first stenosis section on a vessel tree of a first tomographic image;

step A-2, calculating the blood movement distances of the first cross sections at the head and the tail of the inlet section to generate the length of the first inlet section, calculating the blood movement distances of the first cross sections at the head and the tail of the narrowest section to generate the length of the first narrowest section, and calculating the blood movement distances of the first cross sections at the head and the tail of the outlet section to generate the length of the first outlet section;

Step A-3, extracting first area data corresponding to the initial first section of the inlet section to generate first inlet area data, extracting the minimum value in the first area data of all the first sections corresponding to the narrowest section to generate first narrowest area data, and extracting the first area data corresponding to the last first section of the outlet section to generate first outlet area data;

step A-4, calculating the average value of the area data at the first inlet and the area data at the first outlet to generate first inlet and outlet average area data, calculating the ratio of the area data at the first narrowest place to the area data at the first inlet to generate a first narrow area ratio, and calculating the ratio of the area data at the first narrowest place to the average area data at the first inlet and outlet to generate a second narrow area ratio;

step A-5, forming a first narrow section shape feature set from the first inlet section length, the first narrowest section length, the first outlet section length, the first inlet area data, the first narrowest area data and the first outlet area data, and forming a second narrow section shape feature set from the first narrow area ratio and the second narrow area ratio;

step A-6, the first narrow section shape characteristic set and the second narrow section shape characteristic set are combined into first narrow section shape characteristic data;

(2) When analyzing the stenosis location characteristics of the current first stenosis:

b-1, marking the bifurcation points of blood vessels before and after the current first stenosis on a blood vessel tree of the first tomographic image; if the blood vessel bifurcation points exist before and after the current first narrow section, taking the area between the blood vessel bifurcation points closest before and after as a first cut-off area corresponding to the current first narrow section; if the current first narrow section only has the blood vessel bifurcation point in front, taking the area from the nearest blood vessel bifurcation point in front of the current first narrow section to the blood flow outlet of the blood vessel tree as a first interception area corresponding to the current first narrow section; if the current first narrow section only has the blood vessel bifurcation point at the back, taking the area between the blood vessel tree blood flow inlet and the nearest blood vessel bifurcation point at the back of the current first narrow section as a first interception area corresponding to the current first narrow section; if no blood vessel bifurcation exists before and after the current first narrow section, taking all areas from the blood flow inlet to the blood flow outlet of the blood vessel tree as first cut-off areas corresponding to the current first narrow section; calculating the blood movement distance from the initial position of the current first narrow section to the initial position of the first cut-off area, and generating first narrow section position characteristic data;

Step B-2, tracing the crossing point on the vessel tree of the first tomographic image from the initial position of the current first stenosis in the direction opposite to the blood flow movement direction; when tracing back to each crossing point, judging whether the current blood vessel is the main blood vessel at the current crossing point, if not, taking the tracing back path from the current first narrow section to the current crossing point as a first blood vessel area; performing cross point tracing from the end position of the current first narrow section on the blood vessel tree in the same direction as the blood flow moving direction; when tracing back to each crossing point, judging whether the current blood vessel is the main blood vessel at the current crossing point, if not, taking the tracing back path from the current first narrow section to the current crossing point as a first two-blood vessel area; a first vascular region is formed by a first vascular region, a current first narrow section and a first two vascular region; calculating the blood movement distance from the starting position of the current first narrow section to the ending position of the first blood vessel area, and generating first two narrow section position characteristic data;

b-3, tracing back from the initial position of the current first narrow section to the blood flow inlet of the blood vessel tree in the direction opposite to the blood flow moving direction on the blood vessel tree of the first tomographic image, and taking a tracing back path from the current first narrow section to the blood flow inlet of the blood vessel tree as a second blood vessel area; tracing the crossing points on the vessel tree from the end position of the current first narrow section in the same direction as the blood flow moving direction, continuing to trace the crossing points downwards along the main vessel at the current crossing point when tracing each crossing point until reaching the blood flow outlet of the vessel tree, and taking the tracing path from the current first narrow section to the blood flow outlet of the vessel tree as a second vessel region; forming a second blood vessel region from the second first blood vessel region, the current first stenosis and the second blood vessel region; calculating the blood movement distance from the initial position of the current first narrow section to the initial position of the second blood vessel area, and generating first third narrow section position characteristic data;

B-4, combining the first narrow section position characteristic data, the first second narrow section position characteristic data and the first narrow section position characteristic data to form first narrow section position characteristic data;

(3) When analyzing the shape characteristics of the blood vessel where the current first stenosis is located:

c-1, extracting first area data corresponding to the starting position of the first cut-off region to generate first cut-off starting area data, extracting first area data corresponding to a bifurcation point before a current first narrow section in the first cut-off region to generate first cut-off bifurcation area data, extracting minimum first area data in the first cut-off region to generate first cut-off minimum area data, and counting the number of the first narrow sections contained in the first cut-off region to generate first narrow section number; the first blood vessel shape characteristic data is composed of first cut-off initial area data, first cut-off bifurcation area data, first cut-off minimum area data and first narrow section quantity;

step C-2, extracting first area data corresponding to the end position of the first blood vessel region, and generating first two-blood vessel shape characteristic data;

step C-3, counting the number of the first narrow sections from the initial position to the current first narrow section in the second vascular area to generate first third vascular shape characteristic data;

C-4, combining the first blood vessel shape characteristic data, the first second blood vessel shape characteristic data and the first blood vessel shape characteristic data to form first blood vessel shape characteristic data;

step 512, inputting the first stenosis shape feature data, the first stenosis position feature data and the first vessel shape feature data into a training mature stenosis fractional flow reserve variation prediction model to perform stenosis fractional flow reserve variation prediction processing, and generating corresponding first variation data;

here, the stenosis fractional flow reserve variation prediction model may be implemented using a multi-layer Feed-Forward Neural Network (ml_nn) forward neural network, may be implemented using a support vector regression (support vector regression, SVR) model, and may be implemented using a limiting gradient lifting (eXtreme Gradient Boosting, XGBoost) model;

step 52, performing second fractional flow reserve variation prediction processing on each first non-stenosis segment to generate corresponding second variation data;

here, when the second fractional flow reserve variation prediction processing is performed on each first non-stenosis segment, the embodiment of the present invention predicts the fractional flow reserve variation of the current first non-stenosis segment according to the non-stenosis segment shape characteristics of each first non-stenosis segment and the vessel shape characteristics of the vessel in which the first non-stenosis segment is located;

The method specifically comprises the following steps: step 521, analyzing the shape feature of the current first non-narrow segment to generate corresponding first non-narrow segment shape feature data; analyzing the shape characteristics of the blood vessel where the current first non-stenosis section is located, and generating corresponding second blood vessel shape characteristic data;

the first non-narrow segment shape characteristic data is a shape information set of each first non-narrow segment, and the second blood vessel shape characteristic data is an information set of the relative position of the blood vessel segment where each first non-narrow segment is located on the blood vessel tree;

the two analysis processes are specifically as follows:

(1) When analyzing the shape characteristics of the current first non-stenosis:

d-1, calculating blood movement distances of the first section at the head and the tail of the current first non-stenosis section to generate a first non-stenosis section length; extracting first area data corresponding to the initial first section of the current first non-narrow section to generate second inlet area data, and extracting first area data corresponding to the last first section of the current first non-narrow section to generate second outlet area data; calculating the absolute difference value of the area data at the second outlet and the area data at the second inlet to generate first non-narrow section difference value area data; calculating the ratio of the difference value area data of the first non-narrow section to the length of the first non-narrow section to generate a first non-narrow section ratio;

D-2, calculating the curvature radius of the current first non-narrow section, and generating the curvature radius of the first non-narrow section;

d-3, forming first non-narrow section shape characteristic data by the first non-narrow section ratio and the first non-narrow section curvature radius;

(2) When analyzing the shape characteristics of the blood vessel where the current first non-stenosis is located:

e-1, marking the bifurcation points of blood vessels before and after the current first non-stenosis section on a blood vessel tree of the first tomographic image; if the blood vessel bifurcation points exist before and after the current first non-stenosis section, taking the area between the blood vessel bifurcation points closest before and after as a second truncation area corresponding to the current first non-stenosis section; if the current first non-narrow section only has the blood vessel bifurcation point in front, taking the area from the nearest blood vessel bifurcation point in front of the current first non-narrow section to the blood flow outlet of the blood vessel tree as a second interception area corresponding to the current first non-narrow section; if the current first non-narrow section only has the blood vessel bifurcation point at the back, taking the area between the blood vessel tree blood flow inlet and the nearest blood vessel bifurcation point at the back of the current first non-narrow section as a second cut-off area corresponding to the current first non-narrow section; if no blood vessel bifurcation exists before and after the current first non-narrow section, taking all areas from the blood flow inlet to the blood flow outlet of the blood vessel tree as second cut-off areas corresponding to the current first non-narrow section; counting the maximum stenosis rate of all the first stenosis sections contained in the second cut-off region, and generating second blood vessel shape characteristic data;

E-2, tracing the crossing point on the vessel tree of the first tomographic image from the initial position of the current first non-stenosis section in the direction opposite to the blood flow movement direction; when tracing back to each crossing point, judging whether the current blood vessel is the main blood vessel at the current crossing point, if not, taking the tracing back path from the current first non-narrow section to the current crossing point as a third blood vessel area; performing cross point tracing from the end position of the current first non-narrow section on the blood vessel tree in the same direction as the blood flow movement direction; when tracing back to each crossing point, judging whether the current blood vessel is the main blood vessel at the current crossing point, if not, taking the tracing back path from the current first non-narrow section to the current crossing point as a third blood vessel region; forming a third blood vessel region from the third first blood vessel region, the current first non-stenosis and the third second blood vessel region; extracting first area data corresponding to the ending position of the third blood vessel region, and generating second blood vessel shape characteristic data;

e-3, tracing back from the initial position of the current first non-narrow section to the blood flow inlet of the blood vessel tree in the direction opposite to the blood flow moving direction on the blood vessel tree of the first tomographic image, and taking a tracing back path from the current first non-narrow section to the blood flow inlet of the blood vessel tree as a fourth blood vessel region; tracing the crossing points on the vessel tree from the ending position of the current first non-narrow section in the same direction as the blood flow moving direction, continuing to trace the crossing points downwards along the main vessel at the current crossing point when tracing each crossing point until reaching the blood flow outlet of the vessel tree, and taking the tracing path from the current first non-narrow section to the blood flow outlet of the vessel tree as a fourth vessel region; a fourth blood vessel region is formed by the fourth first blood vessel region, the current first non-stenosis section and the fourth second blood vessel region; performing average value calculation on all first area data of a second truncated area closest to the starting position of a fourth blood vessel area to generate first truncated average area data; counting the number of bifurcation points from the starting position of the fourth blood vessel region to the current starting position of the first non-stenosis segment to generate a first bifurcation point number; the first truncated average area data and the first bifurcation point number form second blood vessel shape characteristic data;

E-4, combining the first, second and third blood vessel shape characteristic data to form second blood vessel shape characteristic data;

step 522, inputting the first non-stenosis segment shape characteristic data and the second vessel shape characteristic data into a trained non-stenosis segment fractional flow reserve variation prediction model to perform non-stenosis segment fractional flow reserve variation prediction processing, and generating corresponding second variation data.

Here, the non-stenotic fractional flow reserve variation prediction model may be implemented using an ml_nn network, may be implemented using an SVR model, or may be implemented using an XGBoost model.

Step 6, sequencing the obtained plurality of first variable quantity data and second variable quantity data according to the sequence of each first narrow section and each first non-narrow section to generate a variable quantity data sequence;

wherein the variable data sequence includes a plurality of third variable data.

The sequence of the first stenosis and the first non-stenosis is referred to herein as the sequence starting from the vessel tree inlet in the direction of blood flow.

Step 7, performing fractional flow reserve prediction processing on the cross section fractional flow reserve according to the variable quantity data sequence to generate a first fractional flow reserve data sequence;

Here, since the fractional flow reserve variation of each narrow or non-narrow segment from the inlet on the vessel tree has been obtained, fractional flow reserve values of the head and tail positions of each segment can be obtained sequentially from the first segment; because the blood flow reserve score values of the edge points on the sections of each blood vessel are the same, if the blood flow reserve score values of the edge points of the sections of each section are to be calculated, the blood flow reserve score values of the corresponding sections are only required to be calculated; when calculating the blood flow reserve value of a certain section, calculating according to the blood movement distance between the current section and the head and tail positions of the current section, the blood movement distance between the head and tail positions of the current section and the blood flow reserve value of the head and tail positions of the current section to obtain the blood flow reserve value of the current section= (the blood movement distance between the current section and the head and tail positions of the current section/the blood movement distance between the head and tail positions of the current section) + (the blood movement distance between the current section and the head and tail positions of the current section/the blood movement distance between the head and tail positions of the current section);

The method specifically comprises the following steps: step 71, recording a narrow section or a non-narrow section of the first curve corresponding to the first third variable quantity data as a first curve section; classifying a plurality of first sections corresponding to the first curve segments into a first section set; and storing fractional flow reserve data FFR of a first cross section corresponding to a first one of the first cross section set _1F Setting the first section fractional flow reserve data FFR corresponding to the last first section of the first section set as a preset initial fractional flow reserve _1L Set as (FFR) _1F +first third variation data); designating a first section except the first and last sections in the first section set as a first intermediate section; setting the fractional flow reserve data of the first section corresponding to any first middle section as

Wherein a1 is the distance from the first intermediate section to the first section of the first set of sections, b1 is the distance from the first intermediate section to the last first section of the first set of sections, c1 is the distance from the first to the last first section of the first set of sections;

here, in the first segment, the fractional flow reserve value of the current segment first position, that is, the first fractional flow reserve data FFR of the first section corresponding to the first section of the first section set _1F Reserve fraction for a preset initial blood flow; conventionally the initial fractional flow reserve may be set to 1;

step 72, the narrow section or the non-narrow section of the first curve corresponding to the second and third variable quantity data is recorded as a second curve section; classifying a plurality of first sections corresponding to the second curve segments into a second section set; and storing fractional flow reserve data FFR of a first cross section corresponding to a first cross section of the second cross section set _2F First fractional flow reserve set as corresponding to last first cross section of the first cross section setData FFR _1L First fractional flow reserve data FFR corresponding to the last first cross section of the second set of cross sections _2L Set as (FFR) _2F +second third variation data); marking the first cross section except the first cross section and the last cross section in the second cross section set as a second middle cross section; setting the fractional flow reserve data of the first section corresponding to any second middle section asAnd so on until the last third variance data;

wherein a2 is the distance from the second intermediate section to the first section of the second set of sections, b2 is the distance from the second intermediate section to the last first section of the second set of sections, c2 is the distance from the first to the last first section of the second set of sections;

Here, in the segment following the first segment, the blood flow reserve value of the current segment head position, that is, the blood flow reserve value of the previous segment tail position;

step 73, sorting all the obtained first fractional flow reserve data according to the sequence order, and generating a first fractional flow reserve data sequence.

Here, the first fractional flow reserve data sequence is a set of fractional flow reserve values for all sampled cross sections of the first tomographic image.

If the fractional flow reserve of any vessel position point on the first tomographic image is to be calculated, the two first sections closest in front and back are searched from the first section sequence according to the coordinates of the current vessel position point and the direction of the flow motion: a first front adjacent section and a first rear adjacent section; recording mapping points of the current blood vessel position points on the first front adjacent section and the first rear adjacent section as a first front mapping point and a first rear mapping point; calculating the blood flow movement distance between the first front adjacent section and the first rear adjacent section to generate a first total distance, calculating the blood flow movement distance from the current blood vessel position point to the first front mapping point to generate a first front distance, and calculating the blood flow movement distance from the current blood vessel position point to the first rear mapping point to generate a first rear distance; extracting first fractional flow reserve data corresponding to a first front adjacent section and a first rear adjacent section from the first fractional flow reserve data sequence to generate first front fractional flow reserve data and first rear fractional flow reserve data; calculate fractional flow reserve = (first back pitch first front cross-sectional fractional flow reserve data/first total pitch) + (first front pitch first back cross-sectional fractional flow reserve data/first total pitch) for the current vessel location point.

Fig. 2 is a block diagram of an apparatus for predicting fractional flow reserve according to a blood vessel tomographic image according to a second embodiment of the present invention, where the apparatus may be a terminal device or a server for implementing a method according to an embodiment of the present invention, or may be an apparatus for implementing a method according to an embodiment of the present invention connected to the terminal device or the server, and for example, the apparatus may be an apparatus or a chip system of the terminal device or the server. As shown in fig. 2, the apparatus includes: an acquisition module 201, a data processing module 202, a stenosis and non-stenosis identification module 203, and a fractional flow reserve prediction module 204.

The acquisition module 201 is configured to acquire a first cross-sectional sequence of first tomographic images; the first tomographic image is a three-dimensional tomographic image; the first cross-section sequence includes a plurality of first cross-sections; the first cross section is sampling information obtained by sampling a blood vessel cross section perpendicular to the blood flow direction along the blood flow direction of the first tomographic image.

The data processing module 202 is configured to perform cross-sectional area calculation on each first cross-section to generate corresponding first area data, perform blood movement distance calculation to generate corresponding first distance data, and form a corresponding first area-distance data set from the obtained first area data and the first distance data; and forming a first area-distance data set sequence from the plurality of obtained first area-distance data sets; and performing curve conversion processing on the first area-distance data set sequence by taking the area as an ordinate and the distance as an abscissa to generate a first curve.

The stenosis and non-stenosis identification module 203 is configured to perform a vessel stenosis identification process on the first curve, and generate a plurality of first stenosis; and the curve segments of the first curve separated by respective first narrow segments are denoted as first non-narrow segments.

The fractional flow reserve prediction module 204 is configured to perform a first fractional flow reserve variation prediction process on each first stenosis segment, and generate corresponding first variation data; performing second fractional flow reserve variation prediction processing on each first non-stenosis segment to generate corresponding second variation data; sequencing the obtained plurality of first variable quantity data and second variable quantity data according to the sequence of each first narrow section and each first non-narrow section to generate a variable quantity data sequence; and cross-sectional fractional flow reserve based on the variable data sequence the predictive process generates a first cross-sectional fractional flow reserve data sequence.

The device for predicting fractional flow reserve according to the blood vessel tomography image provided by the embodiment of the invention can execute the method steps in the method embodiment, and the implementation principle and the technical effect are similar, and are not repeated here.

It should be noted that, it should be understood that the division of the modules of the above apparatus is merely a division of a logic function, and may be fully or partially integrated into a physical entity or may be physically separated. And these modules may all be implemented in software in the form of calls by the processing element; or can be realized in hardware; the method can also be realized in a form of calling software by a processing element, and the method can be realized in a form of hardware by a part of modules. For example, the acquisition module may be a processing element that is set up separately, may be implemented in a chip of the above apparatus, or may be stored in a memory of the above apparatus in the form of program code, and may be called by a processing element of the above apparatus and execute the functions of the above determination module. The implementation of the other modules is similar. In addition, all or part of the modules can be integrated together or can be independently implemented. The processing element described herein may be an integrated circuit having signal processing capabilities. In implementation, each step of the above method or each module above may be implemented by an integrated logic circuit of hardware in a processor element or an instruction in a software form.

For example, the modules above may be one or more integrated circuits configured to implement the methods above, such as: 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), etc. For another example, when a module above is implemented in the form of a processing element scheduler code, the processing element may be a general purpose processor, such as a central processing unit (Central Processing Unit, CPU) or other processor that may invoke the program code. For another example, the modules may be integrated together and implemented in the form of a System-on-a-chip (SOC).

In the above embodiments, it may be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented in software, 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 loaded and executed on a computer, produces, in whole or in part, the processes or functions described in accordance with embodiments of the present invention. The computer may be a general purpose computer, a special purpose computer, a computer network, or other programmable apparatus. The 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, from one website, computer, server, or data center via a wired (e.g., coaxial cable, fiber optic, digital subscriber line (Digital Subscriber Line, DSL)) or wireless (e.g., infrared, wireless, bluetooth, microwave, etc.) means. The computer readable storage media may be any available media that can be accessed by a computer or a data storage device such as a server, data center, or the like that contains an integration of one or more available media. The usable medium may be a magnetic medium (e.g., a floppy disk, a hard disk, a magnetic tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a Solid State Disk (SSD)), or the like.

Fig. 3 is a schematic structural diagram of an electronic device according to a third embodiment 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 for implementing the method of the embodiment of the present invention. As shown in fig. 3, the electronic device may include: a processor 301 (e.g., a CPU), a memory 302, a transceiver 303; the transceiver 303 is coupled to the processor 301, and the processor 301 controls the transceiving actions of the transceiver 303. The memory 302 may store various instructions for performing various processing functions and implementing the methods and processes provided in the above-described embodiments of the present invention. Preferably, the electronic device according to the 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 implement communication connections between the elements. The communication port 306 is used for connection communication between the electronic device and other peripheral devices.

The system bus referred to in fig. 3 may be a peripheral component interconnect standard (Peripheral Component Interconnect, PCI) bus, or an extended industry standard architecture (Extended Industry Standard Architecture, EISA) bus, or the like. The system bus may be classified into an address bus, a data bus, a control bus, and the like. For ease of illustration, the figures are shown with only one bold line, but not with only one bus or one type of bus. The communication interface is used to enable communication between the database access apparatus and other devices (e.g., clients, read-write libraries, and read-only libraries). The Memory may comprise 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 processor may be a general-purpose processor, including a Central Processing Unit (CPU), a network processor (Network Processor, NP), etc.; but may also be a digital signal processor DSP, an application specific integrated circuit ASIC, a field programmable gate array FPGA or other programmable logic device, a discrete gate or transistor logic device, a discrete hardware component.

It should be noted that the embodiments of the present invention also provide a computer readable storage medium having instructions stored therein, which when executed on a computer, cause the computer to perform the methods and processes provided in the above embodiments.

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

Those of skill would further appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both, and that the various illustrative elements and steps are described above generally in terms of function in order to clearly illustrate the interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.

The steps of a method or algorithm described in connection with the embodiments disclosed herein may be embodied in hardware, in a software module executed by a processor, or in a combination of the two. The software modules may be disposed in Random Access Memory (RAM), memory, read Only Memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

The foregoing description of the embodiments has been provided for the purpose of illustrating the general principles of the invention, and is not meant to limit the scope of the invention, but to limit the invention to the particular embodiments, and any modifications, equivalents, improvements, etc. that fall within the spirit and principles of the invention are intended to be included within the scope of the invention.

Claims

1. A method of predicting fractional flow reserve from a tomographic image of a blood vessel, the method comprising:

2. The method of predicting fractional flow reserve from a blood vessel tomographic image according to claim 1 wherein said first cross section comprises a plurality of first cross section edge point coordinates; the calculating the cross-sectional area of each first cross-section to generate corresponding first area data, calculating the blood movement distance to generate corresponding first distance data, and forming a corresponding first area-distance data set by the obtained first area data and the first distance data specifically includes:

3. The method of predicting fractional flow reserve based on a tomographic image according to claim 1, wherein said performing a vessel stenosis identification process on said first curve generates a plurality of first stenosis, specifically comprising:

4. A method of predicting fractional flow reserve from a tomographic image according to claim 3 wherein,

Extracting a first area-distance training curve from a preset training dataset before using the first smart model; drawing a two-dimensional blood vessel shape graph according to the first area-distance training curve to generate a first training blood vessel shape graph; performing artificial blood vessel stenosis region marking processing on the first training blood vessel shape graph to generate a first training mark shape graph with a plurality of blood vessel stenosis region marks; inputting the first training blood vessel shape graph into the first intelligent model for first blood vessel stenosis segmentation processing, and generating a second training blood vessel shape graph with a plurality of blood vessel stenosis region marks; performing error calculation on the blood vessel stenosis region marking results of the first training marking shape graph and the second training blood vessel shape graph, and performing reverse modulation on the first intelligent model according to the error calculation result until the error calculation result enters a preset first model error convergence range;

extracting a second area-distance training curve from a preset training dataset before using the second smart model; extracting a three-dimensional tomographic image corresponding to the second area-distance training curve from the training data set, and recording the three-dimensional tomographic image as a first training tomographic image; performing artificial fractional flow reserve calculation on the first training tomographic image according to a computational fluid dynamics method to generate a first training fractional flow reserve data sequence; performing artificial blood vessel stenosis region marking processing on the first training tomographic image according to the first training blood flow reserve fraction data sequence to generate a first training marking tomographic image with a plurality of blood vessel stenosis region marks; inputting the first training tomographic image into the second intelligent model for second blood vessel stenosis segmentation processing, and generating a second training mark tomographic image with a plurality of blood vessel stenosis region marks; and performing error calculation on the blood vessel stenosis region marking results of the first training mark tomography image and the second training mark tomography image, and performing reverse modulation on the second intelligent model according to the error calculation result until the error calculation result enters a preset second model error convergence range.

5. The method for predicting fractional flow reserve based on a tomographic image according to claim 1, wherein said performing a first fractional flow reserve variation prediction process on each of said first stenosis segments generates corresponding first variation data, specifically comprising:

6. The method of predicting fractional flow reserve based on a tomographic image according to claim 1, wherein said performing a second fractional flow reserve variation prediction process on each of said first non-stenosed segments generates corresponding second variation data, specifically comprising:

7. The method of predicting fractional flow reserve from a blood vessel tomographic image according to claim 1, wherein the sequence of delta data comprises a plurality of third delta data; the step of performing fractional flow reserve prediction processing on the cross section fractional flow reserve according to the variable data sequence to generate a first fractional flow reserve data sequence specifically comprises the following steps:

recording a narrow section or a non-narrow section of the first curve corresponding to the first third variable quantity data as a first curve section; classifying a plurality of first sections corresponding to the first curve segments into a first section set; and storing first fractional flow reserve data FFR corresponding to a first one of the first cross-sections of the first set of cross-sections _1F Setting a preset initial Fractional Flow Reserve (FFR) and obtaining first fractional flow reserve data FFR corresponding to the last first section of the first section set _1L Set as (FFR) _1F +first third variation data); recording said first cross-section except for the first and last one of said first set of cross-sections as a first intermediate cross-section; setting the fractional flow reserve data of the first section corresponding to any one of the first middle sections asa1 is the distance from the first intermediate section to the first one of the first set of sections, b1 is the distance from the first intermediate section to the last one of the first set of sections, c1 is the first sectionThe distance from the first to the last of said first cross-sections of the collection;

recording a narrow section or a non-narrow section of the first curve corresponding to the second variable quantity data as a second curve section; classifying a plurality of first sections corresponding to the second curve segments into a second section set; and storing first fractional flow reserve data FFR corresponding to a first one of the first cross-sections of the second set of cross-sections _2F First cross-sectional fractional flow reserve data FFR corresponding to the last of the first cross-sections set of the first cross-sections _1L First cross-sectional fractional flow reserve data FFR corresponding to the last of the first cross-sections of the second set of cross-sections _2L Set as (FFR) _2F +second third variation data); marking the first cross-section except the first and last cross-sections in the second set of cross-sections as a second intermediate cross-section; setting the fractional flow reserve data of the first section corresponding to any one of the second intermediate sections asa2 is the distance from the second intermediate section to the first one of the second set of sections, b2 is the distance from the second intermediate section to the last one of the second set of sections, c2 is the distance from the first one of the second set of sections to the last one of the first sections; and so on until the last of the third variation data;

8. An apparatus for implementing the method steps of predicting fractional flow reserve from a blood vessel tomographic image according to any one of claims 1 to 7, said apparatus comprising: the device comprises an acquisition module, a data processing module, a stenosis and non-stenosis identification module and a fractional flow reserve prediction module;

9. An electronic device, comprising: memory, processor, and transceiver;

the processor being adapted to be coupled to the memory, read and execute the instructions in the memory to implement the method steps of any one of claims 1-7;

10. A computer readable storage medium storing computer instructions which, when executed by a computer, cause the computer to perform the instructions of the method of any one of claims 1-7.