CN115830235B - Three-dimensional model reconstruction method for atrioventricular defect image - Google Patents
Three-dimensional model reconstruction method for atrioventricular defect image Download PDFInfo
- Publication number
- CN115830235B CN115830235B CN202211577782.6A CN202211577782A CN115830235B CN 115830235 B CN115830235 B CN 115830235B CN 202211577782 A CN202211577782 A CN 202211577782A CN 115830235 B CN115830235 B CN 115830235B
- Authority
- CN
- China
- Prior art keywords
- atrioventricular
- defect
- heart
- dimensional
- model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Landscapes
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The invention discloses a three-dimensional model reconstruction method for an atrioventricular defect image, which comprises the following steps: constructing a three-dimensional heart prior model for high-resolution display of the atrioventricular defect part based on a high-resolution heart three-dimensional data set by using 3D slice software; and carrying out model registration on the three-dimensional heart priori model by using the real-time heart atrioventricular defect image to obtain a real-time atrioventricular defect model at the inlet end of the catheter sheath, and marking three-dimensional defect sites in the real-time atrioventricular defect model as three-dimensional movement sites of the inlet end of the catheter sheath. The invention improves the display accuracy of the follow-up real-time heart defect model on the defect part, further ensures that the marked three-dimensional moving site is more accurate, improves the three-dimensional guiding accuracy, utilizes the prior model to perform model registration construction on the heart chamber images obtained in real time to obtain the real-time chamber defect model, does not need all heart images, only utilizes a common initial three-dimensional heart model and a few heart pictures, and has small calculation amount and short reconstruction time.
Description
Technical Field
The invention relates to the technical field of image modeling, in particular to a three-dimensional model reconstruction method for an atrioventricular defect image.
Background
In the occlusion procedure of atrial septal defects of the heart, it is necessary to safely, smoothly and rapidly deliver the occluder to the patient site and to control the occluder to be deployed in both atria to occlude the ostium, respectively.
The general procedure for occlusion surgery is: firstly, a catheter sheath tube passes through a lower vena cava along a peripheral vein (usually a femoral vein) to reach the right atrium of the heart, then, under ultrasound, the position of the sheath tube is observed, the sheath tube is repeatedly tried to run, whether the catheter tube passes through the defect part of an atrial structure (such as an oval foramen and a atrial septal defect) is observed, and then, a plugging umbrella is sent into the relevant defect part along the sheath tube through a guide wire, and the plugging umbrella is sequentially released, so that the plugging umbrella can smoothly clamp the defect.
In the existing operation process, ultrasound guidance is generally adopted, different ultrasound two-dimensional sections are selected, and an experienced sonologist is relied on to search a sheath tube, a guide wire and a plugging umbrella in a matched manner, when a very small atrial defect such as a small oval hole is encountered, the existing ultrasound interface is difficult to correctly display the accurate state of the atrial defect, and in this time, violent operation is often carried out in an operation mode, so that the guide wire or the sheath tube can pass through, but the risk of atrial rupture caused by violent operation is easy to occur, and the existing heart three-dimensional modeling technology can improve the display accuracy of the atrial defect, but has large overall operation amount and long consumption time, is not suitable for guiding the operation in the operation with high timeliness requirement, and therefore, the accurate medical operation is difficult to be achieved in a single ultrasound guiding sheath tube passing mode, and the appearance of an operation area is difficult to intuitively display.
Disclosure of Invention
The invention aims to provide a three-dimensional model reconstruction method for an atrioventricular defect image, which aims to solve the technical problem that three-dimensional reconstruction of the atrioventricular defect with high timeliness and high accuracy is difficult to realize in the prior art.
In order to solve the technical problems, the invention specifically provides the following technical scheme:
a three-dimensional model reconstruction method for an atrioventricular defect image comprises the following steps:
step S1, carrying out resolution improvement on three-dimensional data representing an atrioventricular defect part in a three-dimensional heart data set to obtain a high-resolution heart three-dimensional data set, and constructing a three-dimensional heart priori model for high-resolution display of the atrioventricular defect part based on the high-resolution heart three-dimensional data set by utilizing 3D slice software;
s2, acquiring multi-angle real-time heart atrioventricular images of the periphery of the inlet end of the catheter sheath by using a probe group at the inlet end of the catheter sheath, and performing image recognition on the multi-angle real-time heart atrioventricular images to obtain real-time heart atrioventricular defect images;
and S3, performing model registration on the three-dimensional heart priori model by using a real-time heart atrioventricular defect image to obtain a real-time atrioventricular defect model at the inlet end of the catheter sheath, and marking three-dimensional defect sites in the real-time atrioventricular defect model as three-dimensional movement sites of the inlet end of the catheter sheath so as to realize that the real-time atrioventricular defect model provides real-time three-dimensional guidance for movement of the catheter sheath.
As a preferable scheme of the invention, the method for obtaining the high-resolution heart three-dimensional data set by carrying out resolution improvement on the three-dimensional data representing the atrioventricular defect part in the heart three-dimensional data set comprises the following steps:
carrying out correlation measurement on each three-dimensional data in the heart three-dimensional data set and the atrioventricular defect part to obtain defect data correlation, wherein the correlation measurement formula is as follows:
wherein p is defect data correlation, |S| is data amount of three-dimensional data representing a defective portion of an atrioventricular defect, S is a data set of three-dimensional data representing a defective portion of an atrioventricular defect, f i Three-dimensional data representing the atrioventricular defect site for the ith characterization in S, L is a collection of categories of atrioventricular defect sites, I (f) i L) is f i And L;
performing redundancy measurement on each three-dimensional data in the heart three-dimensional data set to obtain defect data redundancy, wherein the redundancy measurement formula is as follows:
wherein disp is defect data redundancy, S is data quantity of three-dimensional data representing a atrioventricular defect part, S is a data set of three-dimensional data representing the atrioventricular defect part,f i ，f j respectively, the I and j three-dimensional data representing the atrioventricular defect part in S, I and j are count items, I (f) i ,f j ) Is f i And f j Is a mutual information operation formula of (1);
maximizing the correlation of the defect data, minimizing the correlation of the defect data and the redundancy of the defect data, minimizing the redundancy of the defect data, forming an optimization target for screening three-dimensional data representing a chamber defect part in a heart three-dimensional data set, solving the optimization target to obtain three-dimensional data representing a chamber defect part in the heart three-dimensional data set, improving the accuracy of screening the three-dimensional data representing the chamber defect part and reducing the redundancy of the three-dimensional data screening representing the chamber defect part;
the objective function of the optimization objective is as follows: taget=x×max (p) +y×min (disp) ];
wherein Taget is an optimization target value, max and min are respectively a maximizing operator and a minimizing operator, and X, Y is a defect data correlation weight and a defect data redundancy weight;
the functional expression of the defect data correlation weight is as follows:
X＝e -|S| ；
a functional expression of the defect data redundancy weight:
Y＝-e -|S| +1；
performing super-resolution reconstruction on the three-dimensional data representing the atrioventricular defect part by using an image super-resolution reconstruction algorithm to obtain high-resolution three-dimensional data representing the atrioventricular defect part, and replacing the high-resolution three-dimensional data representing the atrioventricular defect part with the high-resolution three-dimensional data representing the atrioventricular defect part in a heart three-dimensional data set to obtain a high-resolution heart three-dimensional data set.
As a preferred scheme of the invention, the construction of the three-dimensional heart priori model for high-resolution display of the atrioventricular defect part based on the high-resolution heart three-dimensional data set by utilizing the 3D slice software comprises the following steps:
the method comprises the steps that high-resolution three-dimensional data representing a atrioventricular defect part and three-dimensional data representing a cavity and an outer wall in a high-resolution heart three-dimensional data set are subjected to part labeling through 3D slice software, and the labeling process comprises cardiovascular;
a three-dimensional cardiac prior model is derived by 3D slicers software, the three-dimensional cardiac prior model comprising six model components of the left atrium, left ventricle, right atrium, right ventricle, left ventricular outer wall and cardiac outer wall.
As a preferred embodiment of the present invention, the method for acquiring multi-angle real-time cardiac atrioventricular images of an outer periphery of an access end of a catheter tube at multiple angles using a probe set at the access end of the catheter tube includes:
according to the heart position coordinate of the inlet end of the catheter sheath tube, and measuring the distance between the heart position coordinate of the inlet end of the catheter sheath tube and the heart position coordinate of the atrioventricular defect part,
if the distance between the heart position coordinate of the inlet end of the catheter sheath tube and the heart position coordinate of the atrioventricular defect part is higher than or equal to a preset threshold value, controlling the quantity of shooting angles of the probe group at the inlet end of the catheter sheath tube to be as follows: n;
if the distance between the heart position coordinate of the inlet end of the catheter sheath tube and the heart position coordinate of the atrioventricular defect part is lower than a preset threshold value, controlling the quantity of shooting angles of the probe group at the inlet end of the catheter sheath tube to be as follows:
wherein N is the basic quantity of shooting angles of the probe group, d is the distance between the heart position coordinate of the inlet end of the catheter sheath tube and the heart position coordinate of the atrioventricular defect part, k is a constant coefficient, and k is more than 2;
the probe group at the inlet end of the catheter sheath tube acquires multi-angle real-time heart atrioventricular images of the periphery of the inlet end of the catheter sheath tube in a shooting angle number, and marks heart position coordinates corresponding to the multi-angle real-time heart atrioventricular images according to heart position coordinates of the inlet end of the catheter sheath tube, wherein the shooting angle number is consistent with the multi-angle real-time heart atrioventricular images.
As a preferable scheme of the invention, the image recognition of the multi-angle real-time heart atrioventricular image to obtain the real-time heart atrioventricular defect image comprises the following steps:
converting the heart position coordinate space of the multi-angle real-time heart atrioventricular image into the inlet end of the catheter sheath tube in the three-dimensional heart prior model, acquiring three-dimensional data corresponding to the multi-angle real-time heart atrioventricular image at the inlet end of the catheter sheath tube in the three-dimensional heart prior model, and comparing the multi-angle real-time heart atrioventricular image with the three-dimensional data corresponding to the multi-angle real-time heart atrioventricular image in a similarity manner, wherein,
if the similarity between the three-dimensional data corresponding to the multi-angle real-time heart atrioventricular image and the multi-angle real-time heart atrioventricular image is higher than the preset similarity, taking the multi-angle real-time heart atrioventricular image corresponding to the similarity as a real-time heart atrioventricular defect image;
if the similarity between the three-dimensional data corresponding to the multi-angle real-time heart atrioventricular image and the multi-angle real-time heart atrioventricular image is lower than or equal to the preset similarity, the multi-angle real-time heart atrioventricular image corresponding to the similarity is used as the non-real-time heart atrioventricular defect image.
As a preferable scheme of the invention, the method for carrying out model registration on the three-dimensional heart priori model by using the real-time heart atrioventricular defect image to obtain the real-time atrioventricular defect model at the inlet end of the catheter sheath comprises the following steps:
and replacing the corresponding three-dimensional data at the inlet end of the catheter sheath in the three-dimensional heart prior model by using the real-time heart atrioventricular defect image, and performing part re-labeling on the three-dimensional data replaced at the inlet end of the catheter sheath in the three-dimensional heart prior model by using 3D slice software, and updating the three-dimensional heart prior model to obtain the real-time atrioventricular defect model at the inlet end of the catheter sheath.
As a preferred embodiment of the present invention, the marking of the three-dimensional defect site in the real-time atrioventricular defect model as the three-dimensional movement site of the entry end of the catheter sheath includes:
inputting a real-time atrioventricular defect model into a defect recognition model pre-established based on a YOLOV5 neural network, and positioning three-dimensional defect sites in the real-time atrioventricular defect model by the defect recognition model;
taking the three-dimensional defect site as a three-dimensional movement site of the inlet end of the catheter sheath;
the defect recognition model is obtained by training a YOLOV5 neural network based on a heart three-dimensional model with a atrioventricular defect part.
As a preferred embodiment of the present invention, the similarity is calculated using euclidean distance.
As a preferable mode of the invention, the probe group at the inlet end of the catheter sheath acquires multi-angle real-time heart atrioventricular images of the periphery of the inlet end of the catheter sheath in a circumferential direction.
As a preferable scheme of the invention, each data component in the heart three-dimensional data set and the multi-angle real-time heart atrioventricular image is normalized.
Compared with the prior art, the invention has the following beneficial effects:
according to the invention, the three-dimensional heart priori model for high-resolution display of the atrioventricular defect part is constructed based on the high-resolution heart three-dimensional data set by utilizing the 3D Slicer software, so that the display accuracy of the follow-up real-time heart defect model on the defect part is improved, the marked three-dimensional moving site is more accurate, the three-dimensional guiding accuracy is improved, the real-time atrioventricular defect model is obtained by carrying out model registration construction on the heart atrioventricular image obtained in real time by utilizing the priori model, the reconstruction of the static three-dimensional heart can be realized by utilizing only one general initial three-dimensional heart model and a few heart pictures without all heart images, the calculated amount is small, the reconstruction time is short, the reconstruction result is accurate and reliable, and the three-dimensional defect site is marked in the real-time atrioventricular defect model as the three-dimensional moving site of the inlet end of the catheter sheath, so that the real-time atrioventricular defect model provides real-time three-time guiding for the movement of the catheter sheath.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below. It will be apparent to those of ordinary skill in the art that the drawings in the following description are exemplary only and that other implementations can be obtained from the extensions of the drawings provided without inventive effort.
FIG. 1 is a flowchart of a three-dimensional model reconstruction method for an atrioventricular defect image according to an embodiment of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but 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, the invention provides a three-dimensional model reconstruction method for an atrioventricular defect image, which comprises the following steps:
step S1, carrying out resolution improvement on three-dimensional data representing an atrioventricular defect part in a three-dimensional heart data set to obtain a high-resolution heart three-dimensional data set, and constructing a three-dimensional heart priori model for high-resolution display of the atrioventricular defect part based on the high-resolution heart three-dimensional data set by utilizing 3D slice software;
the three-dimensional data may be characterized as image three-dimensional data of various parts of the heart.
Performing resolution enhancement on three-dimensional data representing a atrioventricular defect site in a three-dimensional cardiac dataset to obtain a high-resolution three-dimensional cardiac dataset, comprising:
carrying out correlation measurement on each three-dimensional data in the heart three-dimensional data set and the atrioventricular defect part to obtain defect data correlation, wherein a correlation measurement formula is as follows:
wherein p is a deficiencyCorrelation of the notch data, |S| is the data volume of the three-dimensional data representing the atrioventricular defect site, S is the data set of the three-dimensional data representing the atrioventricular defect site, f i Three-dimensional data representing the atrioventricular defect site for the ith characterization in S, L is a collection of categories of atrioventricular defect sites, such as foramen ovale, atrial septal defect, I (f) i L) is f i And L;
performing redundancy measurement on each three-dimensional data in the heart three-dimensional data set to obtain defect data redundancy, wherein a redundancy measurement formula is as follows:
wherein disp is defect data redundancy, |S| is data volume of three-dimensional data representing a atrioventricular defect site, S is a data set of three-dimensional data representing a atrioventricular defect site, and f i ，f j Respectively, the I and j three-dimensional data representing the atrioventricular defect part in S, I and j are count items, I (f) i ,f j ) Is f i And f j Is a mutual information operation formula of (1);
maximizing the correlation of the defect data, minimizing the correlation of the defect data and the redundancy of the defect data, minimizing the redundancy of the defect data, forming an optimization target for screening three-dimensional data representing a chamber defect part in a heart three-dimensional data set, solving the optimization target to obtain three-dimensional data representing a chamber defect part in the heart three-dimensional data set, improving the accuracy of screening the three-dimensional data representing the chamber defect part and reducing the redundancy of the three-dimensional data screening representing the chamber defect part;
the objective function of the optimization objective is: taget=x×max (p) +y×min (disp) ];
wherein Taget is an optimization target value, max and min are respectively a maximizing operator and a minimizing operator, and X, Y is a defect data correlation weight and a defect data redundancy weight;
the functional expression of the defect data correlation weight is:
X＝e -|S| ；
functional expression of defect data redundancy weight:
Y＝-e -|S| +1；
the correlation weight of the defect data is reduced from 1 to 0 along with the increase of the three-dimensional data screened in the S, and the redundancy weight of the defect data is increased from 0 to 1 along with the increase of the three-dimensional data screened in the S, so that the correlation is weighted when the number of the three-dimensional data screened in the early stage of screening is small, and the redundancy is weighted when the number of the three-dimensional data screened in the later stage is large, and the maximum correlation and the minimum redundancy are realized in a self-adaptive mode.
Performing super-resolution reconstruction on the three-dimensional data representing the atrioventricular defect part by using an image super-resolution reconstruction algorithm to obtain high-resolution three-dimensional data representing the atrioventricular defect part, and replacing the high-resolution three-dimensional data representing the atrioventricular defect part with the high-resolution three-dimensional data representing the atrioventricular defect part in a heart three-dimensional data set to obtain a high-resolution heart three-dimensional data set.
The three-dimensional heart prior model is constructed, the three-dimensional heart prior model is used as prior knowledge of atrioventricular defect reconstruction, namely, is used as a reconstruction frame reference of atrioventricular defect reconstruction, the atrioventricular defect reconstruction is carried out on the basis of the three-dimensional heart prior model, reconstruction of the heart model is not needed to be carried out again by utilizing three-dimensional data of all heart parts, reconstruction of the three-dimensional heart model (namely, the real-time atrioventricular defect model) for showing the atrioventricular defect condition at the inlet end of the catheter sheath can be realized by only utilizing one general three-dimensional heart prior model and a few images for representing the atrioventricular defect part at the inlet end of the catheter sheath, the calculated amount is small, the reconstruction time is short, the reconstructed result presents the appearance of an operation area in real time and intuitionally, and the timeliness requirement of operation in operation is met.
When the three-dimensional heart priori model is constructed, each three-dimensional data in the heart three-dimensional data set is screened, the maximum correlation and the minimum redundancy are used as screening objective functions, the correlation between the screened three-dimensional data and the atrioventricular defect position is strongest, the redundancy between the screened three-dimensional data is minimum, invalid representation of the atrioventricular defect position can be avoided, the representativeness of the screened three-dimensional data on the atrioventricular defect position is strongest, finally the screened three-dimensional data can fully represent the atrioventricular defect position, super-resolution reconstruction of the screened three-dimensional data can enable the three-dimensional data representing the atrioventricular defect position to be high in resolution, further the three-dimensional heart priori model constructed can enable the atrioventricular defect position to have accurate expression when three-dimensional reconstruction is carried out, and when the three-dimensional heart defect is reconstructed in a very tiny atrial defect, such as a tiny oval hole, accuracy and display of real-time atrioventricular defect images can be improved, the accuracy and the accuracy of three-dimensional defect position display can be improved, and the three-dimensional defect position guide accuracy can be further improved.
Constructing a three-dimensional heart prior model for high-resolution display of an atrioventricular defect part based on a high-resolution heart three-dimensional data set by using 3D slice software, wherein the three-dimensional heart prior model comprises the following components:
the method comprises the steps that high-resolution three-dimensional data representing a atrioventricular defect part and three-dimensional data representing a cavity and an outer wall in a high-resolution heart three-dimensional data set are subjected to part labeling through 3D slice software, and the labeling process comprises cardiovascular;
a three-dimensional cardiac prior model is derived by 3D Slicer software, the three-dimensional cardiac prior model comprising six model components of the left atrium, left ventricle, right atrium, right ventricle, left ventricular outer wall and cardiac outer wall.
S2, acquiring multi-angle real-time heart atrioventricular images of the periphery of the inlet end of the catheter sheath by using a probe group at the inlet end of the catheter sheath, performing image recognition on the multi-angle real-time heart atrioventricular images to obtain real-time heart atrioventricular defect images, and performing super-resolution reconstruction on the multi-angle real-time heart atrioventricular images by using an image super-resolution reconstruction algorithm to enable the multi-angle real-time heart atrioventricular images to present the same high resolution as a three-dimensional heart prior model;
acquiring multi-angle real-time cardiac atrioventricular images of an outer periphery of an access end of a catheter sheath at multiple angles using a probe set at the access end of the catheter sheath, comprising:
according to the heart position coordinate of the inlet end of the catheter sheath tube, and measuring the distance between the heart position coordinate of the inlet end of the catheter sheath tube and the heart position coordinate of the atrioventricular defect part,
if the distance between the heart position coordinate of the inlet end of the catheter sheath tube and the heart position coordinate of the atrioventricular defect part is higher than or equal to a preset threshold value, controlling the quantity of shooting angles of the probe group at the inlet end of the catheter sheath tube to be as follows: n;
if the distance between the heart position coordinate of the inlet end of the catheter sheath tube and the heart position coordinate of the atrioventricular defect part is lower than a preset threshold value, controlling the quantity of shooting angles of the probe group at the inlet end of the catheter sheath tube to be as follows:
wherein N is the basic quantity of shooting angles of the probe group, d is the distance between the heart position coordinate of the inlet end of the catheter sheath tube and the heart position coordinate of the atrioventricular defect part, k is a constant coefficient, and k is more than 2;
the probe group at the inlet end of the catheter sheath tube acquires multi-angle real-time heart atrioventricular images of the periphery of the inlet end of the catheter sheath tube in a shooting angle number, and marks heart position coordinates corresponding to the multi-angle real-time heart atrioventricular images according to heart position coordinates of the inlet end of the catheter sheath tube, wherein the shooting angle number is consistent with the multi-angle real-time heart atrioventricular images.
The number of shooting angles is determined according to the distance between the position of the inlet end of the catheter and the defect part, when the distance between the position of the inlet end of the catheter and the defect part is smaller, the moving operation of the inlet end of the catheter needs to be finer, more detailed image data are needed to be provided when the real-time atrioventricular defect model is reconstructed, namely, more shooting angles are needed to be provided to obtain more real-time heart atrioventricular defect images, and the real-time atrioventricular defect model is used for displaying the defect part in more detail.
Image recognition is carried out on the multi-angle real-time heart atrioventricular image to obtain a real-time heart atrioventricular defect image, and the method comprises the following steps:
converting the heart position coordinate space of the multi-angle real-time heart atrioventricular image into the inlet end of the catheter sheath tube in the three-dimensional heart prior model, acquiring three-dimensional data corresponding to the multi-angle real-time heart atrioventricular image at the inlet end of the catheter sheath tube in the three-dimensional heart prior model, and comparing the multi-angle real-time heart atrioventricular image with the three-dimensional data corresponding to the multi-angle real-time heart atrioventricular image in a similarity manner, wherein,
if the similarity between the three-dimensional data corresponding to the multi-angle real-time heart atrioventricular image and the multi-angle real-time heart atrioventricular image is higher than the preset similarity, taking the multi-angle real-time heart atrioventricular image corresponding to the similarity as a real-time heart atrioventricular defect image;
if the similarity between the three-dimensional data corresponding to the multi-angle real-time heart atrioventricular image and the multi-angle real-time heart atrioventricular image is lower than or equal to the preset similarity, the multi-angle real-time heart atrioventricular image corresponding to the similarity is used as the non-real-time heart atrioventricular defect image.
And S3, performing model registration on the three-dimensional heart priori model by using a real-time heart atrioventricular defect image to obtain a real-time atrioventricular defect model at the inlet end of the catheter sheath, and marking three-dimensional defect sites in the real-time atrioventricular defect model as three-dimensional movement sites of the inlet end of the catheter sheath so as to realize that the real-time atrioventricular defect model provides real-time three-dimensional guidance for movement of the catheter sheath.
Performing model registration on the three-dimensional heart prior model by using the real-time heart atrioventricular defect image to obtain a real-time atrioventricular defect model at the inlet end of the catheter sheath, wherein the method comprises the following steps:
and replacing the corresponding three-dimensional data at the inlet end of the catheter sheath in the three-dimensional heart prior model by using the real-time heart atrioventricular defect image, and performing part re-labeling on the three-dimensional data replaced at the inlet end of the catheter sheath in the three-dimensional heart prior model by using 3D slice software, and updating the three-dimensional heart prior model to obtain the real-time atrioventricular defect model at the inlet end of the catheter sheath.
Marking a three-dimensional defect site in a real-time atrioventricular defect model as a three-dimensional movement site of an entry end of a catheter sheath, comprising:
inputting a real-time atrioventricular defect model into a defect recognition model pre-established based on a YOLOV5 neural network, and positioning three-dimensional defect sites in the real-time atrioventricular defect model by the defect recognition model;
taking the three-dimensional defect site as a three-dimensional moving site of the inlet end of the catheter sheath;
the defect recognition model is obtained by training a YOLOV5 neural network based on a three-dimensional model of the heart with atrioventricular defect sites.
The heart position coordinates of the catheter sheath inlet end and the heart position coordinates of the real-time heart atrioventricular defect image are converted into a three-dimensional heart priori model to realize the reconstruction of the real-time atrioventricular defect model in the three-dimensional heart priori model, and the YOLOV5 neural network can be used for detecting and positioning the micro target, so that the method is used for identifying the defect and positioning the defect, and further is spatially converted into the heart space coordinates of the catheter sheath so as to guide the movement of the catheter sheath, thereby providing operation assistance for operation in operation and reducing the experience dependence on medical staff.
The similarity is calculated using the euclidean distance.
The probe group at the inlet end of the catheter sheath tube acquires multi-angle real-time heart atrioventricular images of the periphery of the inlet end of the catheter sheath tube in a circumferential direction.
And carrying out normalization processing on each data component in the heart three-dimensional data set and the multi-angle real-time heart atrioventricular image.
Compared with the prior art, the invention has the following beneficial effects:
according to the invention, the three-dimensional heart priori model for high-resolution display of the atrioventricular defect part is constructed based on the high-resolution heart three-dimensional data set by utilizing the 3D Slicer software, so that the display accuracy of the follow-up real-time heart defect model on the defect part is improved, the marked three-dimensional moving site is more accurate, the three-dimensional guiding accuracy is improved, the real-time atrioventricular defect model is obtained by carrying out model registration construction on the heart atrioventricular image obtained in real time by utilizing the priori model, the reconstruction of the static three-dimensional heart can be realized by utilizing only one general initial three-dimensional heart model and a few heart pictures without all heart images, the calculated amount is small, the reconstruction time is short, the reconstruction result is accurate and reliable, and the three-dimensional defect site is marked in the real-time atrioventricular defect model as the three-dimensional moving site of the inlet end of the catheter sheath, so that the real-time atrioventricular defect model provides real-time three-time guiding for the movement of the catheter sheath.
The above embodiments are only exemplary embodiments of the present application and are not intended to limit the present application, the scope of which is defined by the claims. Various modifications and equivalent arrangements may be made to the present application by those skilled in the art, which modifications and equivalents are also considered to be within the scope of the present application.
Claims (9)
1. A three-dimensional model reconstruction method for atrioventricular defect images is characterized by comprising the following steps of: the method comprises the following steps:
step S1, carrying out resolution improvement on three-dimensional data representing an atrioventricular defect part in a three-dimensional heart data set to obtain a high-resolution heart three-dimensional data set, and constructing a three-dimensional heart priori model for high-resolution display of the atrioventricular defect part based on the high-resolution heart three-dimensional data set by utilizing 3D slice software;
s2, acquiring multi-angle real-time heart atrioventricular images of the periphery of the inlet end of the catheter sheath by using a probe group at the inlet end of the catheter sheath, and performing image recognition on the multi-angle real-time heart atrioventricular images to obtain real-time heart atrioventricular defect images;
step S3, performing model registration on the three-dimensional heart priori model by using a real-time heart atrioventricular defect image to obtain a real-time atrioventricular defect model at the inlet end of the catheter sheath, and marking three-dimensional defect sites in the real-time atrioventricular defect model as three-dimensional moving sites at the inlet end of the catheter sheath so as to realize that the real-time atrioventricular defect model provides real-time three-dimensional guidance for the movement of the catheter sheath;
the multi-angle real-time heart atrioventricular image of the periphery of the inlet end of the catheter tube is acquired by utilizing a probe group at the inlet end of the catheter tube at multiple angles, and the multi-angle real-time heart atrioventricular image comprises:
according to the heart position coordinate of the inlet end of the catheter sheath tube, and measuring the distance between the heart position coordinate of the inlet end of the catheter sheath tube and the heart position coordinate of the atrioventricular defect part,
if the distance between the heart position coordinate of the inlet end of the catheter sheath tube and the heart position coordinate of the atrioventricular defect part is higher than or equal to a preset threshold value, controlling the quantity of shooting angles of the probe group at the inlet end of the catheter sheath tube to be as follows: n;
if the distance between the heart position coordinate of the inlet end of the catheter sheath tube and the heart position coordinate of the atrioventricular defect part is lower than a preset threshold value, controlling the quantity of shooting angles of the probe group at the inlet end of the catheter sheath tube to be as follows:
；
wherein N is the basic quantity of shooting angles of the probe group, d is the distance between the heart position coordinate of the inlet end of the catheter sheath tube and the heart position coordinate of the atrioventricular defect part, k is a constant coefficient, and k is more than 2;
the probe group at the inlet end of the catheter sheath tube acquires multi-angle real-time heart atrioventricular images of the periphery of the inlet end of the catheter sheath tube in a shooting angle number, and marks heart position coordinates corresponding to the multi-angle real-time heart atrioventricular images according to heart position coordinates of the inlet end of the catheter sheath tube, wherein the shooting angle number is consistent with the multi-angle real-time heart atrioventricular images.
2. The method for reconstructing a three-dimensional model of an atrioventricular defect image according to claim 1, wherein the method comprises the following steps: the method for obtaining the high-resolution heart three-dimensional data set by improving the resolution of the three-dimensional data representing the atrioventricular defect part in the heart three-dimensional data set comprises the following steps:
carrying out correlation measurement on each three-dimensional data in the heart three-dimensional data set and the atrioventricular defect part to obtain defect data correlation, wherein the correlation measurement formula is as follows:
performing redundancy measurement on each three-dimensional data in the heart three-dimensional data set to obtain defect data redundancy, wherein the redundancy measurement formula is as follows:
wherein disp is defect data redundancy, |S| is data volume of three-dimensional data representing a atrioventricular defect site, S is a data set of three-dimensional data representing a atrioventricular defect site, and f i ，f j Respectively S is the th ini，jThree-dimensional data characterizing the atrioventricular defect site,i，jto count items, I (f i ,f j ) Is f i And f j Is a mutual information operation formula of (1); maximizing the correlation of the defect data, minimizing the correlation of the defect data and the redundancy of the defect data, minimizing the redundancy of the defect data, forming an optimization target for screening three-dimensional data representing a chamber defect part in a heart three-dimensional data set, solving the optimization target to obtain three-dimensional data representing a chamber defect part in the heart three-dimensional data set, improving the accuracy of screening the three-dimensional data representing the chamber defect part and reducing the redundancy of the three-dimensional data screening representing the chamber defect part;
the objective function of the optimization objective is as follows: taget=x×max (p) +y×min (disp) ];
wherein Taget is an optimization target value, max and min are respectively a maximizing operator and a minimizing operator, and X, Y is a defect data correlation weight and a defect data redundancy weight;
the functional expression of the defect data correlation weight is as follows:
a functional expression of the defect data redundancy weight: />
Performing super-resolution reconstruction on the three-dimensional data representing the atrioventricular defect part by using an image super-resolution reconstruction algorithm to obtain high-resolution three-dimensional data representing the atrioventricular defect part, and replacing the high-resolution three-dimensional data representing the atrioventricular defect part with the high-resolution three-dimensional data representing the atrioventricular defect part in a heart three-dimensional data set to obtain a high-resolution heart three-dimensional data set.
3. The method for reconstructing a three-dimensional model of an atrioventricular defect image according to claim 2, wherein the method comprises the following steps: the construction of the three-dimensional heart prior model for high-resolution display of the atrioventricular defect part based on the high-resolution heart three-dimensional data set by utilizing the 3D Slicer software comprises the following steps:
the method comprises the steps that high-resolution three-dimensional data representing a atrioventricular defect part and three-dimensional data representing a cavity and an outer wall in a high-resolution heart three-dimensional data set are subjected to part labeling through 3D slice software, and the labeling process comprises cardiovascular;
a three-dimensional cardiac prior model is derived by 3D slicers software, the three-dimensional cardiac prior model comprising six model components of the left atrium, left ventricle, right atrium, right ventricle, left ventricular outer wall and cardiac outer wall.
4. A method for reconstructing a three-dimensional model of an atrioventricular defect image according to claim 3, wherein: the image recognition is carried out on the multi-angle real-time heart atrioventricular image to obtain a real-time heart atrioventricular defect image, and the method comprises the following steps:
converting the heart position coordinate space of the multi-angle real-time heart atrioventricular image into the inlet end of the catheter sheath tube in the three-dimensional heart prior model, acquiring three-dimensional data corresponding to the multi-angle real-time heart atrioventricular image at the inlet end of the catheter sheath tube in the three-dimensional heart prior model, and comparing the multi-angle real-time heart atrioventricular image with the three-dimensional data corresponding to the multi-angle real-time heart atrioventricular image in a similarity manner, wherein,
if the similarity between the three-dimensional data corresponding to the multi-angle real-time heart atrioventricular image and the multi-angle real-time heart atrioventricular image is higher than the preset similarity, taking the multi-angle real-time heart atrioventricular image corresponding to the similarity as a real-time heart atrioventricular defect image;
if the similarity between the three-dimensional data corresponding to the multi-angle real-time heart atrioventricular image and the multi-angle real-time heart atrioventricular image is lower than or equal to the preset similarity, the multi-angle real-time heart atrioventricular image corresponding to the similarity is used as the non-real-time heart atrioventricular defect image.
5. The method for reconstructing a three-dimensional model of an atrioventricular defect image according to claim 4, wherein the method comprises the steps of: the model registration of the three-dimensional heart prior model by using the real-time heart atrioventricular defect image to obtain a real-time atrioventricular defect model at the inlet end of the catheter sheath comprises the following steps:
and replacing the corresponding three-dimensional data at the inlet end of the catheter sheath in the three-dimensional heart prior model by using the real-time heart atrioventricular defect image, and performing part re-labeling on the three-dimensional data replaced at the inlet end of the catheter sheath in the three-dimensional heart prior model by using 3D slice software, and updating the three-dimensional heart prior model to obtain the real-time atrioventricular defect model at the inlet end of the catheter sheath.
6. The method for reconstructing a three-dimensional model of an atrioventricular defect image according to claim 5, wherein marking a three-dimensional defect site in the real-time atrioventricular defect model as a three-dimensional movement site of an entry end of a catheter sheath comprises:
inputting a real-time atrioventricular defect model into a defect recognition model pre-established based on a YOLOV5 neural network, and positioning three-dimensional defect sites in the real-time atrioventricular defect model by the defect recognition model;
taking the three-dimensional defect site as a three-dimensional movement site of the inlet end of the catheter sheath;
the defect recognition model is obtained by training a YOLOV5 neural network based on a heart three-dimensional model with a atrioventricular defect part.
7. The method for reconstructing a three-dimensional model of an atrioventricular defect image of claim 6, wherein the similarity is calculated using euclidean distance.
8. The method for reconstructing a three-dimensional model of an atrioventricular defect image according to claim 7, wherein the probe set at the entry end of the catheter tube circumferentially acquires multi-angle real-time cardiac atrioventricular images of the outer periphery of the entry end of the catheter tube.
9. The method for reconstructing a three-dimensional model of an atrioventricular defect image according to claim 8, wherein each data component in the three-dimensional dataset of the heart and the multi-angle real-time image of the atrioventricular heart is normalized.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211577782.6A CN115830235B (en) | 2022-12-09 | 2022-12-09 | Three-dimensional model reconstruction method for atrioventricular defect image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211577782.6A CN115830235B (en) | 2022-12-09 | 2022-12-09 | Three-dimensional model reconstruction method for atrioventricular defect image |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115830235A CN115830235A (en) | 2023-03-21 |
| CN115830235B true CN115830235B (en) | 2023-06-27 |
Family
ID=85544783
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211577782.6A Active CN115830235B (en) | 2022-12-09 | 2022-12-09 | Three-dimensional model reconstruction method for atrioventricular defect image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115830235B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1750216A1 (en) * | 2005-08-02 | 2007-02-07 | Biosense Webster, Inc. | Guided procedures for treating atrial fibrillation |
| CN109961508A (en) * | 2019-03-25 | 2019-07-02 | 东南大学 | A kind of heart static three-dimensional model method for reconstructing |
| CN111710043A (en) * | 2020-04-24 | 2020-09-25 | 南通大学附属医院 | Device and method for constructing heart three-dimensional model in real time |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7778686B2 (en) * | 2002-06-04 | 2010-08-17 | General Electric Company | Method and apparatus for medical intervention procedure planning and location and navigation of an intervention tool |
| KR101492801B1 (en) * | 2013-04-17 | 2015-02-12 | 계명대학교 산학협력단 | Operating medical navigation system and method for heart surgery with registration of oct image and 3-dimensional image |
| CN105701013A (en) * | 2016-01-04 | 2016-06-22 | 中国石油大学(华东) | Software defect data feature selection method based on mutual information |
| US10467767B2 (en) * | 2016-12-23 | 2019-11-05 | International Business Machines Corporation | 3D segmentation reconstruction from 2D slices |
| CN107330888A (en) * | 2017-07-11 | 2017-11-07 | 中国人民解放军第三军医大学 | Each chamber dividing method of dynamic heart based on CTA images |
| CN112034789B (en) * | 2020-08-25 | 2021-10-15 | 国家机床质量监督检验中心 | Health assessment method, system and assessment terminal for key parts and complete machine of numerical control machine tool |
| CN112043383B (en) * | 2020-09-30 | 2022-07-15 | 复旦大学附属眼耳鼻喉科医院 | Ophthalmic surgery navigation system and electronic equipment |
| CN112967786B (en) * | 2021-02-26 | 2023-04-18 | 江南大学 | Construction method and system of anatomical navigation based on multimode image and interactive equipment |
-
2022
- 2022-12-09 CN CN202211577782.6A patent/CN115830235B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1750216A1 (en) * | 2005-08-02 | 2007-02-07 | Biosense Webster, Inc. | Guided procedures for treating atrial fibrillation |
| CN109961508A (en) * | 2019-03-25 | 2019-07-02 | 东南大学 | A kind of heart static three-dimensional model method for reconstructing |
| CN111710043A (en) * | 2020-04-24 | 2020-09-25 | 南通大学附属医院 | Device and method for constructing heart three-dimensional model in real time |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115830235A (en) | 2023-03-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| KR102014359B1 (en) | Method and apparatus for providing camera location using surgical video | |
| US9687204B2 (en) | Method and system for registration of ultrasound and physiological models to X-ray fluoroscopic images | |
| RU2653274C2 (en) | Coupled segmentation in conventional and contrast ultrasound 3d images | |
| KR101982149B1 (en) | Method and apparatus for creating medical image using partial medical image | |
| CN102056547B (en) | Medical image processing device and method for processing medical image | |
| US20190130578A1 (en) | Vascular segmentation using fully convolutional and recurrent neural networks | |
| US9155470B2 (en) | Method and system for model based fusion on pre-operative computed tomography and intra-operative fluoroscopy using transesophageal echocardiography | |
| WO2011122035A1 (en) | Projection image generation device, projection image generation programme, and projection image generation method | |
| US20200226752A1 (en) | Apparatus and method for processing medical image | |
| Nurmaini et al. | Accurate detection of septal defects with fetal ultrasonography images using deep learning-based multiclass instance segmentation | |
| EP1851719B1 (en) | Image processing device and method | |
| EP1837828A2 (en) | Image registration using locally-weighted fitting | |
| CN107809955B (en) | Real-time collimation and ROI-filter localization in X-ray imaging via automatic detection of landmarks of interest | |
| EP2505162A1 (en) | Method and apparatus for generating medical image of body organ by using 3-d model | |
| Lekadir et al. | Outlier detection and handling for robust 3-D active shape models search | |
| JP2016525893A (en) | Multimodal segmentation in intravascular images | |
| JP2013516001A (en) | Automatic segmentation and temporal tracking method | |
| JP2006526225A (en) | High speed automatic centerline extraction method and apparatus for virtual endoscope | |
| JP2024029142A (en) | Selecting medical device for use in medical procedure | |
| JP2009519740A (en) | A method for facilitating post-processing of images using deformable meshes | |
| US11847730B2 (en) | Orientation detection in fluoroscopic images | |
| WO2024082441A1 (en) | Deep learning-based multi-modal image registration method and system, and medium | |
| JP5072625B2 (en) | Image processing apparatus and method | |
| EP1851721B1 (en) | A method, a system and a computer program for segmenting a surface in a multidimensional dataset | |
| King et al. | Registering preprocedure volumetric images with intraprocedure 3-D ultrasound using an ultrasound imaging model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |