A method and system for geometric calibration of cardiac CT images based on machine learning

Abstract

The present invention relates to the technical field of image data processing, and specifically discloses a method and system for geometric calibration of cardiac CT images based on machine learning. The method comprises: obtaining a cardiac CT image to be calibrated during atrial contraction, obtaining a basic information data set of an object corresponding to the cardiac CT image to be calibrated, comparing and obtaining a defined characteristic range of geometric calibration of the cardiac CT image during contraction, importing the acquired position characteristic values ​​of the cardiac CT image to be calibrated, the scale characteristic values ​​of the cardiac CT image to be calibrated, and the morphological characteristic values ​​of the cardiac CT image to be calibrated into a geometric calibration model of the cardiac CT image, combining the defined characteristic range of geometric calibration of the cardiac CT image during contraction, and comparing and obtaining a geometric calibration scheme of the cardiac CT image that needs to be geometrically calibrated. The present invention solves the problem that traditional geometric calibration methods often rely on manual marking or simple mathematical models, and the calibration is not comprehensive enough and the accuracy is not high, and is helpful to quickly determine geometric calibration requirements and improve work efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of image data processing, and in particular to a method and system for geometric calibration of cardiac CT images based on machine learning. Background Art

[0002] With the continuous advancement and popularization of computed tomography (CT) technology, cardiac CT imaging technology has become one of the commonly used cardiac examination methods in clinical practice. The research on the geometric calibration method of cardiac CT images is closely related to the optimization of the reconstruction algorithm. Accurate geometric calibration can improve the spatial resolution and geometric accuracy of cardiac CT images, which helps to more accurately evaluate the cardiac structure. Researchers have conducted research on the geometric calibration method of cardiac CT images, aiming to improve the quality and accuracy of cardiac CT images and further improve cardiac imaging diagnostic technology.

[0003] For example, the invention patent with publication number CN117422647A discloses a method and system for geometric calibration of cardiac CT images based on machine learning, which relates to the field of computed tomography (CT) and includes the following steps: building a dynamic simulation experimental platform for the heart in vitro; inputting several different scanning parameters and heart rate parameters into the experimental platform, and obtaining dynamic scanning images of the heart CT as the image database to be calibrated; using a static scan or high-speed camera to scan or photograph the experimental platform, and obtaining a standard image database consistent with the dynamic scanning image of the heart CT; constructing a geometric calibration model of the heart CT image based on machine learning, and sending the image database to be calibrated and the standard image database into the geometric calibration model of the heart CT for training and verification; obtaining the trained geometric calibration model of the heart CT image, and inputting the individualized dynamic scanning image of the heart CT of the patient to obtain the calibrated real CT image. The present invention can realize the rapid and accurate calibration of the dynamic scanning CT image of the heart.

[0004] Nowadays, there are still some shortcomings in the research on a method for geometric calibration of cardiac CT images based on machine learning. Specifically, traditional geometric calibration methods often rely on manual labeling or simple mathematical models, which are difficult to adapt to complex cardiac structures, and the calibration is not comprehensive enough and the accuracy is not high. Manual labeling of structures in cardiac CT images requires a lot of time and effort. There may be subjective differences in the labeling results between different labelers, resulting in inconsistent and unreliable data. Traditional simple mathematical models usually assume that the cardiac structure has a certain regular shape, which is difficult to cope with the complex geometric morphology and individual differences of the heart, and cannot fully calibrate the overall geometric morphology of the heart. Summary of the invention

[0005] In view of the deficiencies in the prior art, the present invention provides a method and system for geometric calibration of cardiac CT images based on machine learning, which can effectively solve the problems involved in the above-mentioned background technology.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: In a first aspect, the present invention provides a method for geometric calibration of cardiac CT images based on machine learning, comprising the following steps: in atrial contraction, obtaining a cardiac CT image to be calibrated, obtaining a basic information data set of objects corresponding to the cardiac CT image to be calibrated, and obtaining a geometric calibration definition feature range of the cardiac CT image in the contraction period based on the obtained basic information data set of objects corresponding to the cardiac CT image to be calibrated; obtaining a position data set of the cardiac CT image to be calibrated, and obtaining a position feature value of the cardiac CT image to be calibrated through comprehensive analysis based on the obtained position data set of the cardiac CT image to be calibrated; obtaining a scale data set of the cardiac CT image to be calibrated, and obtaining a scale feature value of the cardiac CT image to be calibrated through comprehensive analysis based on the obtained scale data set of the cardiac CT image to be calibrated. CT image scale feature value; obtaining a morphological data set of the heart CT image to be calibrated, the morphological data set of the heart CT image to be calibrated specifically includes the heart wall thickness of the heart CT image to be calibrated, the area size of the left ventricle chamber of the heart CT image to be calibrated, the area size of the right ventricle chamber of the heart CT image to be calibrated, and the diameter of the coronary artery of the heart CT image to be calibrated; based on the obtained morphological data set of the heart CT image to be calibrated, a comprehensive analysis is performed to obtain the morphological feature value of the heart CT image to be calibrated; the position feature value of the heart CT image to be calibrated, the scale feature value of the heart CT image to be calibrated, and the morphological feature value of the heart CT image to be calibrated are introduced into the heart CT image geometric calibration model, the feature range is defined in combination with the geometric calibration of the heart CT image during systole, and the geometric calibration scheme of the heart CT image that needs to be geometrically calibrated is compared.

[0007] As a further method, the basic information data set of the object corresponding to the cardiac CT image to be calibrated specifically includes the age of the object corresponding to the cardiac CT image to be calibrated, the gender of the object corresponding to the cardiac CT image to be calibrated, and the BMI of the object corresponding to the cardiac CT image to be calibrated.

[0008] As a further method, based on the acquired basic information data set corresponding to the object of the cardiac CT image to be calibrated, the geometric calibration feature range of the systolic cardiac CT image is obtained by comparison. The specific analysis process is: generating a designated label for the age of the object corresponding to the cardiac CT image to be calibrated, the gender of the object corresponding to the cardiac CT image to be calibrated, and the BMI of the object corresponding to the cardiac CT image to be calibrated; comparing the designated label with the geometric calibration feature range of the systolic cardiac CT image corresponding to each designated label stored in the database, to obtain the geometric calibration feature range of the systolic cardiac CT image corresponding to the designated label.

[0009] As a further method, based on the acquired position data set of the heart CT image to be calibrated, a comprehensive analysis is performed to obtain the position characteristic value of the heart CT image to be calibrated. The specific analysis process is: the position data set of the heart CT image to be calibrated is acquired, and the position data set of the heart CT image to be calibrated specifically includes the straight-line distance between the left atrium and the left ventricle of the heart CT image to be calibrated, the straight-line distance between the right atrium and the right ventricle of the heart CT image to be calibrated, and the straight-line distance between the left atrium and the right atrium of the heart CT image to be calibrated; based on the acquired position data set of the heart CT image to be calibrated, a comprehensive analysis is performed to obtain the position characteristic value of the heart CT image to be calibrated, and the position characteristic value of the heart CT image to be calibrated is used as an analysis basis for determining whether the heart CT image to be calibrated needs to be geometrically calibrated.

[0010] As a further method, the position feature value of the heart CT image is to be calibrated, and the specific analysis process is as follows:

[0011] ;

[0012] In the formula, is the position feature value of the cardiac CT image to be calibrated, is the straight-line distance between the left atrium and left ventricle of the heart CT image to be calibrated, is the straight-line distance between the right atrium and the right ventricle in the heart CT image to be calibrated, is the straight-line distance between the left atrium and the right atrium of the heart CT image to be calibrated, is the compensation factor of the straight-line distance between the left atrium and the left ventricle of the cardiac CT image to be calibrated, is the compensation factor of the straight-line distance between the right atrium and the right ventricle of the cardiac CT image to be calibrated, is the compensation factor of the straight-line distance between the left atrium and the right atrium of the cardiac CT image to be calibrated, and e is a natural constant.

[0013] As a further method, based on the acquired scale data set of the heart CT image to be calibrated, a comprehensive analysis is performed to obtain the scale characteristic value of the heart CT image to be calibrated. The specific analysis process is: the scale data set of the heart CT image to be calibrated is acquired, and the scale data set of the heart CT image to be calibrated specifically includes the area covered by the heart front view of the heart CT image to be calibrated, the area covered by the heart right view of the heart CT image to be calibrated, and the ratio of the left atrium to the right atrium coverage area of ​​the heart CT image to be calibrated; based on the acquired scale data set of the heart CT image to be calibrated, a comprehensive analysis is performed to obtain the scale characteristic value of the heart CT image to be calibrated, and the scale characteristic value of the heart CT image to be calibrated is used as an analysis basis for determining whether the heart CT image to be calibrated needs to be geometrically calibrated.

[0014] As a further method, the scale feature value of the cardiac CT image is to be calibrated, and the specific analysis process is as follows:

[0015] ;

[0016] In the formula, is the scale feature value of the cardiac CT image to be calibrated, is the area covered by the heart front view of the heart CT image to be calibrated, is the area covered by the right heart view of the cardiac CT image to be calibrated, is the coverage area ratio of the left atrium and right atrium of the cardiac CT image to be calibrated, is the compensation factor for the area covered by the cardiac front view of the cardiac CT image to be calibrated, is the compensation factor for the area covered by the right cardiac view of the cardiac CT image to be calibrated, is the compensation factor for the ratio of the left atrium to the right atrium coverage area of ​​the cardiac CT image to be calibrated.

[0017] As a further method, based on the acquired morphological data set of the cardiac CT images to be calibrated, a comprehensive analysis is performed to obtain the morphological feature values ​​of the cardiac CT images to be calibrated. The specific analysis process is: based on the acquired morphological data set of the cardiac CT images to be calibrated, a comprehensive analysis is performed to obtain the morphological feature values ​​of the cardiac CT images to be calibrated. The morphological feature values ​​of the cardiac CT images to be calibrated are used as an analysis basis for determining whether the cardiac CT images to be calibrated need to be geometrically calibrated.

[0018] As a further method, a geometric calibration scheme of the cardiac CT image that needs to be geometrically calibrated is compared, and the specific analysis process is: the position characteristic value of the cardiac CT image to be calibrated, the scale characteristic value of the cardiac CT image to be calibrated, and the morphological characteristic value of the cardiac CT image to be calibrated are imported into the cardiac CT image geometric calibration model, and the geometric calibration characteristic value of the cardiac CT image is obtained by comprehensive analysis. The geometric calibration characteristic value of the cardiac CT image is used as an analysis basis for judging whether the cardiac CT image to be calibrated needs to be geometrically calibrated; the geometric calibration characteristic value of the cardiac CT image is compared with the characteristic range defined by the geometric calibration of the systolic cardiac CT image; if the geometric calibration characteristic value of the cardiac CT image belongs to the characteristic range defined by the geometric calibration of the systolic cardiac CT image, then the cardiac CT image to be calibrated corresponding to the geometric calibration characteristic value of the cardiac CT image does not need to be geometrically calibrated. Calibration; if the geometric calibration characteristic value of the cardiac CT image does not belong to the characteristic range defined by the geometric calibration of the systolic cardiac CT image, then the cardiac CT image to be calibrated corresponding to the geometric calibration characteristic value of the cardiac CT image needs to be geometrically calibrated, and the difference between the median of the characteristic range defined by the geometric calibration of the systolic cardiac CT image and the geometric calibration characteristic value of the cardiac CT image is recorded as the geometric calibration deviation value of the cardiac CT image, and the geometric calibration schemes corresponding to the geometric calibration deviation values ​​of each cardiac CT image stored in the database are compared to obtain the geometric calibration scheme corresponding to the geometric calibration deviation value of the cardiac CT image, and the geometric calibration scheme corresponding to the geometric calibration deviation value of the cardiac CT image represents the geometric calibration scheme of the cardiac CT image that needs to be geometrically calibrated corresponding to the geometric calibration deviation value of the cardiac CT image.

[0019] The second aspect of the present invention provides a machine learning-based cardiac CT image geometric calibration system, comprising a geometric calibration defined feature range comparison module, a position feature value acquisition module, a scale feature value acquisition module, a morphological feature value acquisition module and a geometric calibration scheme comparison module, wherein: the geometric calibration defined feature range comparison module is used to obtain a cardiac CT image to be calibrated during atrial contraction, obtain a basic information data set of an object corresponding to the cardiac CT image to be calibrated, and compare and obtain a defined feature range of the systolic cardiac CT image based on the obtained basic information data set of the object corresponding to the cardiac CT image to be calibrated; the position feature value acquisition module is used to obtain a position data set of the cardiac CT image to be calibrated, and obtain a position data set of the cardiac CT image to be calibrated based on the obtained position data set of the cardiac CT image to be calibrated by comprehensive analysis. The scale characteristic value acquisition module is used to acquire the scale data set of the heart CT image to be calibrated, and obtain the scale characteristic value of the heart CT image to be calibrated by comprehensive analysis based on the acquired scale data set of the heart CT image to be calibrated; the morphological characteristic value acquisition module is used to acquire the morphological data set of the heart CT image to be calibrated, and obtain the morphological characteristic value of the heart CT image to be calibrated by comprehensive analysis based on the acquired morphological data set of the heart CT image to be calibrated; the geometric calibration scheme comparison module is used to import the position characteristic value of the heart CT image to be calibrated, the scale characteristic value of the heart CT image to be calibrated and the morphological characteristic value of the heart CT image to be calibrated into the geometric calibration model of the heart CT image, combine the geometric calibration of the heart CT image in the systolic period to define the characteristic range, and compare the geometric calibration scheme of the heart CT image that needs to be geometrically calibrated.

[0020] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects:

[0021] (1) The present invention provides a method and system for geometric calibration of cardiac CT images based on machine learning. According to the basic information data set of the object corresponding to the cardiac CT image to be calibrated, combined with characteristic values ​​such as position, scale and morphology, a targeted geometric calibration plan can be formulated. Customized calibration can be performed according to individual conditions to improve the calibration effect. By automatically comparing and analyzing the characteristic values ​​of the cardiac CT image to be calibrated, the geometric calibration requirements can be quickly determined and a corresponding calibration plan can be formulated. The automated process can save time and labor costs, improve work efficiency, reduce errors and improve the consistency and reliability of calibration.

[0022] (2) The present invention obtains the cardiac CT image to be calibrated during atrial contraction, obtains the basic information data set of the object corresponding to the cardiac CT image to be calibrated, and compares to obtain the geometric calibration feature range of the cardiac CT image during contraction, thereby ensuring that the heart is in a certain physiological state, which is conducive to accurately capturing the morphological and structural characteristics of the heart. This helps to determine the feature range of geometric calibration, improve the accuracy of calibration, provide a clear physiological reference standard, and more accurately compare to obtain the geometric calibration feature range of the cardiac CT image. Acquiring the cardiac CT image during atrial contraction can reduce the influence of motion artifacts, improve the clarity and stability of the image, and help to accurately capture the geometric characteristics of the heart, determine the range of geometric calibration, and avoid calibration errors caused by motion.

[0023] (3) The present invention introduces the position characteristic values ​​of the cardiac CT image to be calibrated, the scale characteristic values ​​of the cardiac CT image to be calibrated and the morphological characteristic values ​​of the cardiac CT image to be calibrated into the geometric calibration model of the cardiac CT image. The geometric calibration model of the cardiac CT image is combined with the geometric calibration of the cardiac CT image during systole to define the characteristic range, and a geometric calibration scheme of the cardiac CT image that needs to be geometrically calibrated is compared. A targeted calibration scheme can be formulated to ensure the precision and accuracy of the calibration, which helps to reduce image distortion and deformation, and improve the reliability and accuracy of the cardiac CT image. The geometric calibration scheme obtained by automatic comparison and analysis can reduce the influence of human subjective factors on the geometric calibration results, which helps to reduce errors and improve the consistency and reliability of the calibration. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The present invention is further described using the accompanying drawings, but the embodiments in the accompanying drawings do not constitute any limitation to the present invention. A person skilled in the art can obtain other drawings based on the following drawings without creative work.

[0025] Figure 1 The figure is a schematic flow chart of the method steps of the present invention.

[0026] Figure 2 It is a schematic diagram of system module connection of the present invention.

[0027] Figure 3 It is an image in which the geometric calibration feature values ​​of the cardiac CT image change with the position feature values ​​of the cardiac CT image to be calibrated. DETAILED DESCRIPTION

[0028] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0029] Reference Figure 1 As shown, the first aspect of the present invention provides a method for geometric calibration of cardiac CT images based on machine learning, including: during atrial contraction, obtaining a cardiac CT image to be calibrated, obtaining a basic information data set of objects corresponding to the cardiac CT image to be calibrated, and based on the obtained basic information data set of objects corresponding to the cardiac CT image to be calibrated, comparing to obtain a defined feature range for the geometric calibration of the cardiac CT image during contraction.

[0030] Specifically, the basic information data set of the object corresponding to the cardiac CT image to be calibrated includes the age of the object corresponding to the cardiac CT image to be calibrated, the gender of the object corresponding to the cardiac CT image to be calibrated, and the BMI of the object corresponding to the cardiac CT image to be calibrated. The age of the object corresponding to the cardiac CT image to be calibrated refers to the age of the individual who receives the cardiac CT image scan, and the age can provide important physiological information in the analysis of the cardiac CT image. The gender of the object corresponding to the cardiac CT image to be calibrated indicates the gender of the individual who receives the cardiac CT image scan, that is, male or female. Gender may affect differences in cardiac structure and function in the analysis of cardiac CT images. The BMI of the object corresponding to the cardiac CT image to be calibrated, BMI is the abbreviation of body mass index, which is an indicator to measure the relationship between a person's weight and height. BMI can help evaluate a person's weight status, and thus provide a certain reference for the analysis of cardiac CT images.

[0031] In a specific embodiment, as people age, their metabolism slows down, leading to weight gain or higher BMI. There are physiological differences between men and women in terms of body fat content, muscle mass, etc., which leads to different body fat content between men and women at the same BMI. Gender may affect the interpretation and assessment of BMI. There are also some physiological differences between age and gender. As people age, men and women may show different characteristics in terms of heart structure and function.

[0032] Specifically, based on the acquired basic information data set corresponding to the object of the cardiac CT image to be calibrated, the geometric calibration feature range of the systolic cardiac CT image is obtained by comparison. The specific analysis process is: generating a designated label for the age of the object corresponding to the cardiac CT image to be calibrated, the gender of the object corresponding to the cardiac CT image to be calibrated, and the BMI of the object corresponding to the cardiac CT image to be calibrated; comparing the designated label with the geometric calibration feature range of the systolic cardiac CT image corresponding to each designated label stored in the database, to obtain the geometric calibration feature range of the systolic cardiac CT image corresponding to the designated label.

[0033] It needs to be explained that the above-mentioned generation of designated labels based on individual information such as age, gender and BMI of the subject can realize personalized calibration of cardiac CT images. Subjects of different ages, genders and BMIs may have differences in cardiac structure and function. Therefore, personalized calibration can more accurately adapt to the characteristics of different subjects. Comparing the designated labels with the characteristic ranges defined by the geometric calibration of systolic cardiac CT images corresponding to each designated label stored in the database can achieve more accurate calibration. The compared characteristic ranges can more accurately locate and calibrate the cardiac CT image, thereby improving the accuracy and reliability of calibration. Calibration is one of the key steps in cardiac CT image analysis. Accurate calibration can ensure that subsequent analysis results are accurate. Generating designated labels based on the individual information of the subject and comparing the corresponding characteristic ranges can improve the calibration quality of cardiac CT images, thereby improving accuracy.

[0034] It needs to be explained that the above-mentioned method of obtaining the cardiac CT image to be calibrated during atrial contraction, obtaining the basic information data set of the object corresponding to the cardiac CT image to be calibrated, and comparing the defined feature range of the geometric calibration of the cardiac CT image during contraction can ensure that the heart is in a certain physiological state, which is conducive to accurately capturing the morphological and structural characteristics of the heart, which helps to determine the defined feature range of the geometric calibration, improve the accuracy of the calibration, provide a clear physiological reference standard, and more accurately compare the defined feature range of the geometric calibration of the cardiac CT image. Acquiring the cardiac CT image during atrial contraction can reduce the impact of motion artifacts, improve the clarity and stability of the image, and help to accurately capture the geometric features of the heart, determine the range of geometric calibration, and avoid calibration errors caused by motion.

[0035] A data set of cardiac CT image positions to be calibrated is obtained, and based on the obtained data set of cardiac CT image positions to be calibrated, a comprehensive analysis is performed to obtain a feature value of the cardiac CT image positions to be calibrated.

[0036] Specifically, based on the acquired position data set of the heart CT image to be calibrated, a comprehensive analysis is performed to obtain the position feature value of the heart CT image to be calibrated. The specific analysis process is: obtaining the position data set of the heart CT image to be calibrated, the position data set of the heart CT image to be calibrated specifically includes the straight-line distance between the left atrium and the left ventricle of the heart CT image to be calibrated, the straight-line distance between the right atrium and the right ventricle of the heart CT image to be calibrated, and the straight-line distance between the left atrium and the right atrium of the heart CT image to be calibrated. The straight-line distance between the left atrium and the left ventricle of the heart CT image to be calibrated, the straight-line distance between the right atrium and the right ventricle of the heart CT image to be calibrated, and the straight-line distance between the left atrium and the right atrium of the heart CT image to be calibrated are obtained through a heart CT scanner. Heart CT scanning is a non-invasive imaging technique that can obtain three-dimensional images of the heart with high resolution to display the structure and function of the heart. The user lies in the scanner and, through specific contrast agents or different scanning modes, can obtain parameters such as the shape, size, and distance of each chamber inside the heart, including the straight-line distance between the left atrium and the left ventricle, the straight-line distance between the right atrium and the right ventricle, and the straight-line distance between the left atrium and the right atrium. Based on the acquired data set of the position of the heart CT image to be calibrated, a comprehensive analysis is performed to obtain the position characteristic values ​​of the heart CT image to be calibrated, which are used as the analysis basis for determining whether the heart CT image to be calibrated needs geometric calibration.

[0037] It needs to be explained that when the left ventricle of the heart enlarges, the straight-line distance between the left atrium and the left ventricle may increase. When the load on the right atrium increases, the straight-line distance between the right atrium and the right ventricle may increase. The straight-line distance between the left atrium and the right atrium will also be affected by the structure of the heart and may change to a certain extent.

[0038] It needs to be explained that the above-mentioned comprehensive analysis of position data such as the straight-line distance between the left atrium and the left ventricle, the straight-line distance between the right atrium and the right ventricle, and the straight-line distance between the left atrium and the right atrium can obtain more comprehensive cardiac structure information and reflect the geometric relationship between different chambers inside the heart. The position characteristic value is a quantitative analysis through specific data, which helps to provide objective evaluation indicators and can more accurately describe the characteristics of the cardiac structure. The position characteristic value can provide standardized evaluation indicators and help establish a reference range or standard to compare data from different objects and different time points, discover abnormalities or changes, and improve the accuracy and consistency of the results. The position characteristic value serves as an analytical basis for determining whether the cardiac CT image to be calibrated needs geometric calibration. If the position characteristic value shows that there is an obvious deviation or abnormality in the position of the cardiac structure, geometric calibration may be required to obtain more accurate results.

[0039] Furthermore, the specific analysis process of the position characteristic value of the heart CT image to be calibrated is as follows:

[0040] ;

[0041] In the formula, is the position feature value of the cardiac CT image to be calibrated, is the straight-line distance between the left atrium and left ventricle of the heart CT image to be calibrated, is the straight-line distance between the right atrium and the right ventricle in the heart CT image to be calibrated, is the straight-line distance between the left atrium and the right atrium of the heart CT image to be calibrated, is the compensation factor of the straight-line distance between the left atrium and the left ventricle of the cardiac CT image to be calibrated, is the compensation factor of the straight-line distance between the right atrium and the right ventricle of the cardiac CT image to be calibrated, is the compensation factor of the straight-line distance between the left atrium and the right atrium of the cardiac CT image to be calibrated, and e is a natural constant.

[0042] It should be explained that the above-mentioned position characteristic value of the cardiac CT image to be calibrated is calculated by the straight-line distance between the left atrium and the left ventricle of the cardiac CT image to be calibrated, the straight-line distance between the right atrium and the right ventricle of the cardiac CT image to be calibrated, and the straight-line distance between the left atrium and the right atrium of the cardiac CT image to be calibrated. The straight-line distance between the left atrium and the left ventricle of the cardiac CT image to be calibrated, the straight-line distance between the right atrium and the right ventricle of the cardiac CT image to be calibrated, and the straight-line distance between the left atrium and the right atrium of the cardiac CT image to be calibrated are normalized, so that the position relationship of the left atrium, the left ventricle, the right atrium and the right ventricle in the cardiac CT image can be accurately located. The straight-line distance between the left atrium and the left ventricle, the straight-line distance between the right atrium and the right ventricle, and the straight-line distance between the left atrium and the right atrium reflect the relative position and morphological characteristics of the internal structure of the heart, which can evaluate the morphology and size of the heart and help to analyze heart lesions more accurately. The calculation of the position characteristic value can be used as auxiliary information to help identify whether geometric calibration of the cardiac CT image is needed. If the position characteristic value deviates greatly from the standard value, it may indicate that the image needs to be calibrated to improve the geometric accuracy. By understanding the straight-line distance between the left and right atria and chambers, the display ratio of the image can be better adjusted. , which helps to present clearer and more accurate cardiac CT images, thereby improving image quality. The compensation factors of the straight-line distance between the left atrium and left ventricle of the cardiac CT image to be calibrated, the straight-line distance between the right atrium and right ventricle of the cardiac CT image to be calibrated, and the straight-line distance between the left atrium and right atrium of the cardiac CT image to be calibrated are obtained from the database. The straight-line distance between the left atrium and left ventricle of the cardiac CT image to be calibrated, the straight-line distance between the right atrium and right ventricle of the cardiac CT image to be calibrated, and the straight-line distance between the left atrium and right atrium of the cardiac CT image to be calibrated and the straight-line distance between the left atrium and right atrium of the cardiac CT image to be calibrated are established based on historical data. A mapping set of compensation factors of the atrium-left ventricle straight-line distance, the right atrium-right ventricle straight-line distance of the heart CT image to be calibrated, and the left atrium-right atrium straight-line distance of the heart CT image to be calibrated is obtained, and the compensation factors of the left atrium-left ventricle straight-line distance of the heart CT image to be calibrated, the right atrium-right ventricle straight-line distance of the heart CT image to be calibrated, and the left atrium-right atrium straight-line distance of the heart CT image to be calibrated corresponding to the current left atrium-left ventricle straight-line distance of the heart CT image to be calibrated, the right atrium-right ventricle straight-line distance of the heart CT image to be calibrated, and the left atrium-right atrium straight-line distance of the heart CT image to be calibrated are obtained.

[0043] A scale dataset of the cardiac CT image to be calibrated is obtained, and based on the obtained scale dataset of the cardiac CT image to be calibrated, a scale feature value of the cardiac CT image to be calibrated is obtained through comprehensive analysis.

[0044] Specifically, based on the acquired scale data set of the heart CT image to be calibrated, a comprehensive analysis is performed to obtain the scale feature value of the heart CT image to be calibrated. The specific analysis process is: obtaining the scale data set of the heart CT image to be calibrated, the scale data set of the heart CT image to be calibrated specifically includes the area covered by the front view of the heart of the heart CT image to be calibrated, the area covered by the right view of the heart of the heart CT image to be calibrated, and the ratio of the left atrium to the right atrium coverage area of ​​the heart CT image to be calibrated. The area covered by the front view of the heart of the heart CT image to be calibrated: this refers to the area covered by the front view angle of the heart on the CT image (observing the heart from the front), which can reflect the size of the range displayed by the heart in the front view, and provides important information for evaluating the overall morphology and structure of the heart. The area covered by the right view of the heart of the heart CT image to be calibrated: this refers to the area covered by the right view angle of the heart on the CT image (observing the heart from the side). Evaluating the size of the range displayed by the heart in the right view is helpful to understand the structure and position of the heart on the side. The ratio of the left atrium to the right atrium coverage area of ​​the heart CT image to be calibrated: this refers to the area covered by the left atrium and right atrium of the heart on the CT image. The ratio reflects the relative area relationship between the left atrium and the right atrium on the CT image, and provides a reference for evaluating the proportions and proportional relationships of different chambers inside the heart. The area covered by the front view of the heart of the heart CT image to be calibrated, the area covered by the right view of the heart of the heart CT image to be calibrated, and the ratio of the left atrium to the right atrium covered area of ​​the heart CT image to be calibrated are obtained through a heart CT scanner. Heart CT scanning is a non-invasive imaging examination technology that can obtain three-dimensional images of the heart with high resolution to show the structure and function of the heart. The user lies down in the scanner and uses specific contrast agents or different scanning modes to obtain parameters such as the shape, size and distance of each chamber inside the heart, including the area covered by the front view of the heart of the heart CT image to be calibrated, the area covered by the right view of the heart of the heart CT image to be calibrated, and the ratio of the left atrium to the right atrium covered area of ​​the heart CT image to be calibrated. Based on the acquired scale data set of the heart CT image to be calibrated, the scale feature value of the heart CT image to be calibrated is obtained by comprehensive analysis, and the scale feature value of the heart CT image to be calibrated is used as an analysis basis for judging whether the heart CT image to be calibrated needs to be geometrically calibrated.

[0045] It needs to be explained that there may be a correlation between the areas of the heart displayed on the front view and the right view. When the area of ​​the heart displayed on the front view is larger, a larger area may also be displayed on the right view. The area relationship between the front view and the right view can help to more comprehensively evaluate the overall morphology and three-dimensional structure of the heart. The coverage area ratio of the left atrium to the right atrium of the heart can reflect the relative size ratio of the left atrium and the right atrium on the CT image, and may also have a certain relationship with the coverage area ratio of the front view and the right view. It can more clearly understand the proportion of the left atrium and the right atrium in the heart structure, and further understand the structure and proportional relationship of the chambers inside the heart.

[0046] It should be explained that the eigenvalues ​​obtained based on the scale data set are objective and quantitative evaluation indicators, which can provide standardized and consistent analysis methods. They can be quantitatively evaluated based on the specific data of the eigenvalues, avoid subjective judgment and give geometric calibration suggestions, and improve the objectivity of geometric calibration decisions. By comprehensively analyzing multiple parameters such as the area covered by the front view and right view of the heart and the ratio of the left atrium to the right atrium coverage area, the scale eigenvalues ​​can be obtained to more comprehensively reflect the geometric characteristics of the cardiac CT image, which is helpful to more accurately judge the geometric calibration of the image, thereby improving the accuracy of the evaluation. With the help of the analysis of the scale data set and the eigenvalues, it is possible to more quickly evaluate whether the cardiac CT image to be calibrated needs geometric calibration, and make decisions based on the eigenvalue quantitative indicators, which can save time, improve work efficiency, and avoid subjective judgment and repeated operations.

[0047] Furthermore, the specific analysis process of the cardiac CT image scale feature value to be calibrated is as follows:

[0048] ;

[0049] In the formula, is the scale feature value of the cardiac CT image to be calibrated, is the area covered by the heart front view of the heart CT image to be calibrated, is the area covered by the right heart view of the cardiac CT image to be calibrated, is the coverage area ratio of the left atrium and right atrium of the cardiac CT image to be calibrated, is the compensation factor for the area covered by the cardiac front view of the cardiac CT image to be calibrated, is the compensation factor for the area covered by the right cardiac view of the cardiac CT image to be calibrated, is the compensation factor for the ratio of the left atrium to the right atrium coverage area of ​​the cardiac CT image to be calibrated.

[0050] It should be explained that the scale characteristic value of the cardiac CT image to be calibrated is calculated by the area covered by the front view of the heart of the cardiac CT image to be calibrated, the area covered by the right view of the heart of the cardiac CT image to be calibrated, and the ratio of the left atrium to the right atrium covered by the cardiac CT image to be calibrated. The area covered by the front view of the heart of the cardiac CT image to be calibrated, the area covered by the right view of the heart of the cardiac CT image to be calibrated, and the ratio of the left atrium to the right atrium covered by the cardiac CT image to be calibrated are normalized to calculate the scale characteristic value of the cardiac CT image to be calibrated, which can evaluate the scale of the cardiac CT image to be calibrated, and is helpful to calibrate the image. The ratio of the left atrium to the right atrium reflects the relative size relationship between the left atrium and the right atrium in the image, and can evaluate the relative area relationship between the left and right atria in the image, so as to better understand the characteristics of the cardiac structure. The calculation of the scale eigenvalue can be used as auxiliary information to help identify whether geometric calibration of cardiac CT images is needed. If the area ratio of the area covered by the front view and the right view of the heart is significantly different, or the area ratio of the left atrium to the right atrium is far from the standard value, it may indicate that the image needs to be calibrated to adjust Scale and proportion, through the analysis of scale eigenvalues, the scale proportion of the image can be better adjusted to improve the display quality and accuracy of the image, the set area covered by the heart front view of the heart CT image to be calibrated, the area covered by the heart right view of the heart CT image to be calibrated, and the compensation factor of the ratio of the left atrium to the right atrium coverage area of ​​the heart CT image to be calibrated are obtained from the database, and the historically measured area covered by the heart front view of the heart CT image to be calibrated, the area covered by the heart right view of the heart CT image to be calibrated, and the ratio of the left atrium to the right atrium coverage area of ​​the heart CT image to be calibrated are established according to historical data A mapping set of compensation factors of the area covered by the front view of the heart of the CT image, the area covered by the right view of the heart of the heart to be calibrated, and the ratio of the left atrium to the right atrium coverage area of ​​the heart CT image to be calibrated is obtained, and the compensation factors of the area covered by the front view of the heart of the heart CT image to be calibrated, the area covered by the right view of the heart of the heart CT image to be calibrated, and the ratio of the left atrium to the right atrium coverage area of ​​the heart CT image to be calibrated are obtained.

[0051] A morphological data set of a cardiac CT image to be calibrated is obtained, and based on the obtained morphological data set of the cardiac CT image to be calibrated, a morphological feature value of the cardiac CT image to be calibrated is obtained through comprehensive analysis.

[0052] Specifically, based on the acquired morphological data set of the heart CT image to be calibrated, a comprehensive analysis is performed to obtain the morphological feature value of the heart CT image to be calibrated. The specific analysis process is: the morphological data set of the heart CT image to be calibrated is obtained, and the morphological data set of the heart CT image to be calibrated specifically includes the heart wall thickness of the heart CT image to be calibrated, the left ventricular chamber area of ​​the heart CT image to be calibrated, the right ventricular chamber area of ​​the heart CT image to be calibrated, and the coronary artery diameter of the heart CT image to be calibrated. The heart wall thickness refers to the thickness of the contractile part of the heart muscle, which can reflect the contraction and relaxation function of the heart. The left ventricle is the main pump of the heart, responsible for pumping oxygenated blood to the whole body. The area of ​​the left ventricular chamber can reflect the size and volume of the left ventricle. The right ventricle receives venous blood from the systemic circulation and sends it to the pulmonary circulation for oxygenation. The area of ​​the right ventricular chamber can reflect the size and volume of the right ventricle. The coronary artery is the main blood supply artery of the heart, which directly affects the blood supply to the myocardium. The wall thickness, the area size of the left ventricular chamber of the heart CT image to be calibrated, the area size of the right ventricular chamber of the heart CT image to be calibrated, and the diameter of the coronary artery of the heart CT image to be calibrated are obtained through a heart CT scanner. Heart CT scanning is a non-invasive imaging examination technology that can obtain three-dimensional images of the heart with high resolution to show the structure and function of the heart. The user lies in the scanner and uses specific contrast agents or different scanning modes to obtain parameters such as the morphology, size, and distance of each chamber inside the heart, including the heart wall thickness of the heart CT image to be calibrated, the area size of the left ventricular chamber of the heart CT image to be calibrated, the area size of the right ventricular chamber of the heart CT image to be calibrated, and the diameter of the coronary artery of the heart CT image to be calibrated; based on the acquired morphological data set of the heart CT image to be calibrated, a comprehensive analysis is performed to obtain the morphological feature values ​​of the heart CT image to be calibrated, and the morphological feature values ​​of the heart CT image to be calibrated are used as the analysis basis for judging whether the heart CT image to be calibrated needs to be geometrically calibrated.

[0053] In a specific embodiment, the thickness of the heart wall is generally related to the size and function of the left ventricle. An increase in the left ventricular chamber area may lead to thickening of the heart wall. The function of the right ventricle is affected by the structure and thickness of the myocardium. An increase in the thickness of the heart wall may affect the diastolic and systolic functions of the right ventricle, thereby causing the size of the right ventricular chamber area to change. There is an interaction between the left ventricle and the right ventricle. The ratio between the size of the left ventricular chamber area and the size of the right ventricular chamber area can reflect the balance between the ventricles and the overall functional state of the heart. The coronary artery is the main blood supply artery of the heart, and its diameter can affect the blood supply to the myocardium. Changes in the thickness of the heart wall and the size of the left and right ventricular chamber areas may affect the blood perfusion state of the coronary arteries.

[0054] It should be explained that the above analysis of morphological parameters such as heart wall thickness, left ventricular chamber area, right ventricular chamber area and coronary artery diameter can realize quantitative evaluation of heart structure and function, and the morphological feature values ​​obtained by comprehensive analysis of the morphological data set of heart CT images to be calibrated can help doctors have a more comprehensive understanding of the morphological characteristics of the heart.

[0055] It should be explained that the specific analysis process of the morphological feature values ​​of the cardiac CT image to be calibrated is as follows:

[0056] ;

[0057] In the formula, is the morphological feature value of the cardiac CT image to be calibrated, is the heart wall thickness of the heart CT image to be calibrated, is the size of the left ventricular chamber area of ​​the cardiac CT image to be calibrated, is the size of the right ventricular chamber in the cardiac CT image to be calibrated, is the coronary artery diameter of the cardiac CT image to be calibrated, is the compensation factor of the heart wall thickness of the cardiac CT image to be calibrated, is the compensation factor of the left ventricular chamber area of ​​the cardiac CT image to be calibrated, is the compensation factor of the right ventricular chamber area of ​​the cardiac CT image to be calibrated, is the compensation factor of the coronary artery diameter of the cardiac CT image to be calibrated, and e is a natural constant.

[0058] It should be explained that the above-mentioned morphological feature values ​​of the cardiac CT image to be calibrated are calculated by the cardiac wall thickness of the cardiac CT image to be calibrated, the area of ​​the left ventricle chamber of the cardiac CT image to be calibrated, the area of ​​the right ventricle chamber of the cardiac CT image to be calibrated, and the diameter of the coronary artery of the cardiac CT image to be calibrated. The cardiac wall thickness of the cardiac CT image to be calibrated, the area of ​​the left ventricle chamber of the cardiac CT image to be calibrated, the area of ​​the right ventricle chamber of the cardiac CT image to be calibrated, and the diameter of the coronary artery of the cardiac CT image to be calibrated are normalized. The cardiac wall thickness, the area of ​​the left and right ventricles and the diameter of the coronary artery are important indicators for evaluating the structure and function of the heart, and comprehensively evaluate the anatomical structure and function of the heart. Morphological features help to better understand the overall condition of the heart. By analyzing the morphological feature values, especially comparing indicators such as heart wall thickness, left and right ventricular chamber area, and coronary artery diameter, the geometric correction of the image can be detected. If there is a geometric calibration problem in the image, it may cause structural size distortion. The analysis of morphological feature values ​​can help to timely discover and correct geometric calibration problems. Geometric calibration problems will cause image proportion distortion, making it impossible to truly reflect structural features. By calculating the morphological feature values, it can be detected whether the geometric features of the image are in line with the actual proportion, thereby adjusting the proportion and scale of the image and improving the image quality. Geometric calibration problems will affect different CT images. Comparison and analysis between them can be performed to determine whether the morphological feature values ​​meet the standards. The images can be standardized so that the images taken by different patients and different equipment are comparable. The compensation factors of the heart wall thickness of the cardiac CT image to be calibrated, the left ventricular cavity area of ​​the cardiac CT image to be calibrated, the right ventricular cavity area of ​​the cardiac CT image to be calibrated, and the coronary artery diameter of the cardiac CT image to be calibrated are obtained from the database. The historically measured heart wall thickness of the cardiac CT image to be calibrated, the left ventricular cavity area of ​​the cardiac CT image to be calibrated, the right ventricular cavity area of ​​the cardiac CT image to be calibrated, and the coronary artery diameter of the cardiac CT image to be calibrated are established based on historical data. A mapping set of compensation factors of the heart wall thickness of the cardiac CT image to be calibrated, the left ventricular chamber area of ​​the cardiac CT image to be calibrated, the right ventricular chamber area of ​​the cardiac CT image to be calibrated, and the coronary artery diameter of the cardiac CT image to be calibrated is obtained to obtain the compensation factors of the heart wall thickness of the cardiac CT image to be calibrated, the left ventricular chamber area of ​​the cardiac CT image to be calibrated, the right ventricular chamber area of ​​the cardiac CT image to be calibrated, and the coronary artery diameter of the cardiac CT image to be calibrated corresponding to the current heart wall thickness of the cardiac CT image to be calibrated, the left ventricular chamber area of ​​the cardiac CT image to be calibrated, the right ventricular chamber area of ​​the cardiac CT image to be calibrated, and the coronary artery diameter of the cardiac CT image to be calibrated.

[0059] The position feature values, scale feature values ​​and morphological feature values ​​of the cardiac CT image to be calibrated are imported into the cardiac CT image geometric calibration model, and the feature range is defined in combination with the geometric calibration of the cardiac CT image during systole, and the geometric calibration scheme of the cardiac CT image that needs to be geometrically calibrated is compared.

[0060] Specifically, a geometric calibration scheme of a cardiac CT image that needs to be geometrically calibrated is compared, and the specific analysis process is: the position characteristic value of the cardiac CT image to be calibrated, the scale characteristic value of the cardiac CT image to be calibrated, and the morphological characteristic value of the cardiac CT image to be calibrated are imported into the geometric calibration model of the cardiac CT image, and the geometric calibration characteristic value of the cardiac CT image is obtained through comprehensive analysis. The geometric calibration characteristic value of the cardiac CT image is used as an analysis basis for judging whether the cardiac CT image to be calibrated needs to be geometrically calibrated; the geometric calibration characteristic value of the cardiac CT image is compared with the characteristic range defined by the geometric calibration of the systolic cardiac CT image; if the geometric calibration characteristic value of the cardiac CT image belongs to the characteristic range defined by the geometric calibration of the systolic cardiac CT image, then the cardiac CT image to be calibrated corresponding to the geometric calibration characteristic value of the cardiac CT image does not need to be geometrically calibrated. If the geometric calibration characteristic value of the cardiac CT image does not belong to the characteristic range defined by the geometric calibration of the systolic cardiac CT image, then the cardiac CT image to be calibrated corresponding to the geometric calibration characteristic value of the cardiac CT image needs to be geometrically calibrated, and the difference between the median of the characteristic range defined by the geometric calibration of the systolic cardiac CT image and the geometric calibration characteristic value of the cardiac CT image is recorded as the geometric calibration deviation value of the cardiac CT image, and the geometric calibration deviation value of the cardiac CT image is compared with the geometric calibration schemes corresponding to the geometric calibration deviation values ​​of each cardiac CT image stored in the database to obtain the geometric calibration scheme corresponding to the geometric calibration deviation value of the cardiac CT image, and the geometric calibration scheme corresponding to the geometric calibration deviation value of the cardiac CT image indicates the geometric calibration scheme of the cardiac CT image that needs to be geometrically calibrated corresponding to the geometric calibration deviation value of the cardiac CT image.

[0061] In a specific embodiment, by comprehensively analyzing the positional eigenvalues, scale eigenvalues ​​and morphological eigenvalues ​​of the cardiac CT image and determining whether geometric calibration is needed based on the corresponding geometric calibration model, accurate geometric calibration of the cardiac CT image to be calibrated can be achieved to ensure the data accuracy and quality of the image. The automated geometric calibration process can reduce the interference of human errors, ensure the accuracy and consistency of geometric calibration, and improve work efficiency and reliability. The corresponding geometric calibration scheme can be obtained based on the geometric calibration deviation value of the cardiac CT image, and a personalized geometric calibration strategy can be implemented. Accurate calibration adjustments can be made according to the specific situation of each cardiac CT image to be calibrated, thereby improving the effect of geometric calibration.

[0062] It should be explained that the specific analysis process of the above cardiac CT image geometric calibration model is as follows:

[0063] ;

[0064] In the formula, is the geometric calibration eigenvalue of cardiac CT images, is the position feature value of the cardiac CT image to be calibrated, is the scale feature value of the cardiac CT image to be calibrated, is the morphological characteristic value of the cardiac CT image to be calibrated, and e is a natural constant.

[0065] It needs to be explained that, by importing the position characteristic values ​​of the cardiac CT image to be calibrated, the scale characteristic values ​​of the cardiac CT image to be calibrated and the morphological characteristic values ​​of the cardiac CT image to be calibrated into the cardiac CT image geometric calibration model, the cardiac CT image geometric calibration model, combined with the geometric calibration of the cardiac CT image during systole, defines the characteristic range, and compares the geometric calibration scheme of the cardiac CT image that needs to be geometrically calibrated. A targeted calibration scheme can be formulated to ensure the precision and accuracy of the calibration, help reduce image distortion and deformation, and improve the reliability and accuracy of the cardiac CT image. The geometric calibration scheme obtained by automatic comparison and analysis can reduce the influence of human subjective factors on the geometric calibration results, help reduce errors and improve the consistency and reliability of the calibration.

[0066] like Figure 3 As shown, it is an image of the change of the geometric calibration eigenvalue of the cardiac CT image as the position eigenvalue of the cardiac CT image to be calibrated changes, wherein the x-axis represents the position eigenvalue of the cardiac CT image to be calibrated, and the y-axis represents the geometric calibration eigenvalue of the cardiac CT image, which can help intuitively understand how the position eigenvalue of the cardiac CT image to be calibrated affects the geometric calibration eigenvalue of the cardiac CT image. The larger the position eigenvalue of the cardiac CT image to be calibrated, the larger the geometric calibration eigenvalue of the cardiac CT image, indicating that the quality of the cardiac CT image to be calibrated is better. As the position eigenvalue of the cardiac CT image to be calibrated increases, the influence of the position eigenvalue of the cardiac CT image to be calibrated on the geometric calibration eigenvalue of the cardiac CT image gradually weakens. The scale eigenvalue of the cardiac CT image to be calibrated is set to 1 and remains unchanged, and the morphological eigenvalue of the cardiac CT image to be calibrated is set to 2 and remains unchanged, and only the size of the position eigenvalue of the cardiac CT image to be calibrated is changed. Example values ​​of the position eigenvalue of the cardiac CT image to be calibrated are as follows:

[0067] Table 1: Example values ​​of cardiac CT image position feature values ​​to be calibrated in the cardiac CT image geometric calibration feature values

[0068]

[0069] It should be noted that the larger the scale eigenvalue of the cardiac CT image to be calibrated, the larger the geometric calibration eigenvalue of the cardiac CT image, indicating that the quality of the cardiac CT image to be calibrated is better. As the scale eigenvalue of the cardiac CT image to be calibrated increases, the influence of the scale eigenvalue of the cardiac CT image to be calibrated on the geometric calibration eigenvalue of the cardiac CT image gradually weakens. The position eigenvalue of the cardiac CT image to be calibrated is set to 2 and remains unchanged. The morphological eigenvalue of the cardiac CT image to be calibrated is set to 2 and remains unchanged. Only the size of the scale eigenvalue of the cardiac CT image to be calibrated is changed. Example values ​​of the scale eigenvalue of the cardiac CT image to be calibrated are as follows:

[0070] Table 2: Example values ​​of the cardiac CT image scale feature values ​​to be calibrated in the cardiac CT image geometric calibration feature values

[0071]

[0072] Reference Figure 2 As shown, the second aspect of the present invention provides a cardiac CT image geometric calibration system based on machine learning, including a geometric calibration definition feature range comparison module, a position feature value acquisition module, a scale feature value acquisition module, a morphological feature value acquisition module and a geometric calibration scheme comparison module.

[0073] The geometric calibration defined feature range comparison module is used to obtain the cardiac CT image to be calibrated during atrial contraction, obtain the basic information data set of the object corresponding to the cardiac CT image to be calibrated, and compare and obtain the geometric calibration defined feature range of the cardiac CT image during contraction based on the obtained basic information data set of the object corresponding to the cardiac CT image to be calibrated.

[0074] The position feature value acquisition module is used to acquire the position data set of the heart CT image to be calibrated, and based on the acquired position data set of the heart CT image to be calibrated, comprehensively analyze and obtain the position feature value of the heart CT image to be calibrated.

[0075] The scale feature value acquisition module is used to acquire the scale data set of the cardiac CT image to be calibrated, and based on the acquired scale data set of the cardiac CT image to be calibrated, comprehensively analyze and obtain the scale feature value of the cardiac CT image to be calibrated.

[0076] The morphological feature value acquisition module is used to acquire the morphological data set of the heart CT image to be calibrated, and based on the acquired morphological data set of the heart CT image to be calibrated, comprehensively analyze and obtain the morphological feature value of the heart CT image to be calibrated.

[0077] The geometric calibration scheme comparison module is used to import the position characteristic values ​​of the cardiac CT image to be calibrated, the scale characteristic values ​​of the cardiac CT image to be calibrated and the morphological characteristic values ​​of the cardiac CT image to be calibrated into the geometric calibration model of the cardiac CT image, combine the geometric calibration of the systolic cardiac CT image to define the characteristic range, and compare the geometric calibration scheme of the cardiac CT image that needs to be geometrically calibrated.

[0078] It needs to be explained that the above-mentioned method and system for geometric calibration of cardiac CT images based on machine learning can formulate a targeted geometric calibration plan according to the basic information data set of the object corresponding to the cardiac CT image to be calibrated, combined with characteristic values ​​such as position, scale and morphology, and can perform customized calibration according to individual conditions to improve the calibration effect. By automatically comparing and analyzing the characteristic values ​​of the cardiac CT image to be calibrated, the geometric calibration requirements can be quickly determined and a corresponding calibration plan can be formulated. The automated process can save time and labor costs, improve work efficiency, and reduce errors and improve the consistency and reliability of calibration.

[0079] The above contents are merely examples and explanations of the structure of the present invention. The technicians in this technical field may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the structure of the invention or exceed the scope defined by the claims, they should all fall within the protection scope of the present invention.

Claims

1. A method for geometric calibration of cardiac CT images based on machine learning, characterized in that: The following steps are involved: During atrial contraction, a cardiac CT image to be calibrated is obtained, a basic information dataset of an object corresponding to the cardiac CT image to be calibrated is obtained, and a geometric calibration definition feature range of the cardiac CT image during contraction is obtained by comparison based on the obtained basic information dataset of the object corresponding to the cardiac CT image to be calibrated; Acquire a data set of cardiac CT image positions to be calibrated, and obtain a position feature value of the cardiac CT image to be calibrated by comprehensive analysis based on the acquired data set of cardiac CT image positions to be calibrated; Acquire a scale dataset of the cardiac CT image to be calibrated, and obtain a scale feature value of the cardiac CT image to be calibrated by comprehensive analysis based on the acquired scale dataset of the cardiac CT image to be calibrated; Acquire a morphological data set of a heart CT image to be calibrated, wherein the morphological data set of the heart CT image to be calibrated specifically includes the heart wall thickness of the heart CT image to be calibrated, the area size of the left ventricle chamber of the heart CT image to be calibrated, the area size of the right ventricle chamber of the heart CT image to be calibrated, and the diameter of the coronary artery of the heart CT image to be calibrated; based on the acquired morphological data set of the heart CT image to be calibrated, comprehensively analyze and obtain the morphological feature value of the heart CT image to be calibrated; Importing the position feature value, scale feature value and morphology feature value of the cardiac CT image to be calibrated into the cardiac CT image geometric calibration model, combining the systolic cardiac CT image geometric calibration to define the feature range, and comparing the geometric calibration scheme of the cardiac CT image that needs geometric calibration; The specific analysis process of comparing the geometric calibration scheme of the cardiac CT image that needs to be geometrically calibrated is as follows: Importing the position characteristic value of the cardiac CT image to be calibrated, the scale characteristic value of the cardiac CT image to be calibrated, and the morphological characteristic value of the cardiac CT image to be calibrated into the cardiac CT image geometric calibration model, and comprehensively analyzing to obtain the geometric calibration characteristic value of the cardiac CT image, and the geometric calibration characteristic value of the cardiac CT image is used as an analysis basis for judging whether the cardiac CT image to be calibrated needs to be geometrically calibrated; Comparing the geometric calibration characteristic value of the cardiac CT image with the geometric calibration defined characteristic range of the systolic cardiac CT image; If the geometric calibration feature value of the cardiac CT image belongs to the defined feature range of the geometric calibration of the cardiac CT image in the systolic period, then the to-be-calibrated cardiac CT image corresponding to the geometric calibration feature value of the cardiac CT image does not need to be geometrically calibrated; If the geometric calibration characteristic value of the cardiac CT image does not belong to the characteristic range defined by the geometric calibration of the systolic cardiac CT image, then the cardiac CT image to be calibrated corresponding to the geometric calibration characteristic value of the cardiac CT image needs to be geometrically calibrated, and the difference between the median of the characteristic range defined by the geometric calibration of the systolic cardiac CT image and the geometric calibration characteristic value of the cardiac CT image is recorded as the geometric calibration deviation value of the cardiac CT image, and the geometric calibration deviation value of the cardiac CT image is compared with the geometric calibration schemes corresponding to the geometric calibration deviation values ​​of each cardiac CT image stored in the database to obtain the geometric calibration scheme corresponding to the geometric calibration deviation value of the cardiac CT image, and the geometric calibration scheme corresponding to the geometric calibration deviation value of the cardiac CT image represents the geometric calibration scheme of the cardiac CT image that needs to be geometrically calibrated corresponding to the geometric calibration deviation value of the cardiac CT image.

2. The method for geometric calibration of cardiac CT images based on machine learning according to claim 1, characterized in that: The basic information data set of the object corresponding to the cardiac CT image to be calibrated specifically includes the age of the object corresponding to the cardiac CT image to be calibrated, the gender of the object corresponding to the cardiac CT image to be calibrated, and the BMI of the object corresponding to the cardiac CT image to be calibrated.

3. The method for geometric calibration of cardiac CT images based on machine learning according to claim 2, characterized in that: The basic information data set corresponding to the object of the cardiac CT image to be calibrated is obtained, and the geometric calibration feature range of the systolic cardiac CT image is obtained by comparison. The specific analysis process is as follows: Generate a designated label for the age of the subject corresponding to the cardiac CT image to be calibrated, the gender of the subject corresponding to the cardiac CT image to be calibrated, and the BMI of the subject corresponding to the cardiac CT image to be calibrated; The designated tag is compared with the systolic cardiac CT image geometric calibration defined characteristic ranges corresponding to the designated tags stored in the database to obtain the systolic cardiac CT image geometric calibration defined characteristic ranges corresponding to the designated tag.

4. The method for geometric calibration of cardiac CT images based on machine learning according to claim 1, characterized in that: Based on the acquired data set of cardiac CT image position to be calibrated, the position feature value of the cardiac CT image to be calibrated is obtained by comprehensive analysis. The specific analysis process is as follows: Acquire a heart CT image position data set to be calibrated, the heart CT image position data set to be calibrated specifically including a straight-line distance between the left atrium and the left ventricle of the heart CT image to be calibrated, a straight-line distance between the right atrium and the right ventricle of the heart CT image to be calibrated, and a straight-line distance between the left atrium and the right atrium of the heart CT image to be calibrated; Based on the acquired data set of cardiac CT image position to be calibrated, a comprehensive analysis is performed to obtain a position characteristic value of the cardiac CT image to be calibrated, and the position characteristic value of the cardiac CT image to be calibrated is used as an analysis basis for determining whether the cardiac CT image to be calibrated needs geometric calibration.

5. The method for geometric calibration of cardiac CT images based on machine learning according to claim 4, characterized in that: The specific analysis process of the position characteristic value of the cardiac CT image to be calibrated is as follows: ； In the formula, is the position feature value of the cardiac CT image to be calibrated, is the straight-line distance between the left atrium and left ventricle of the heart CT image to be calibrated, is the straight-line distance between the right atrium and the right ventricle in the heart CT image to be calibrated, is the straight-line distance between the left atrium and the right atrium of the heart CT image to be calibrated, is the compensation factor of the straight-line distance between the left atrium and the left ventricle of the cardiac CT image to be calibrated, is the compensation factor of the straight-line distance between the right atrium and the right ventricle of the cardiac CT image to be calibrated, is the compensation factor of the straight-line distance between the left atrium and the right atrium of the cardiac CT image to be calibrated, and e is a natural constant.

6. The method for geometric calibration of cardiac CT images based on machine learning according to claim 1, characterized in that: Based on the acquired cardiac CT image scale data set to be calibrated, the scale feature value of the cardiac CT image to be calibrated is obtained by comprehensive analysis. The specific analysis process is as follows: Acquire a scale data set of the cardiac CT image to be calibrated, wherein the scale data set of the cardiac CT image to be calibrated specifically includes an area covered by a front view of the heart of the cardiac CT image to be calibrated, an area covered by a right view of the heart of the cardiac CT image to be calibrated, and a ratio of a left atrium to a right atrium coverage area of ​​the cardiac CT image to be calibrated; Based on the acquired scale data set of the cardiac CT image to be calibrated, a comprehensive analysis is performed to obtain the scale feature value of the cardiac CT image to be calibrated, and the scale feature value of the cardiac CT image to be calibrated is used as an analysis basis for determining whether the cardiac CT image to be calibrated needs to be geometrically calibrated.

7. The method for geometric calibration of cardiac CT images based on machine learning according to claim 6, characterized in that: The specific analysis process of the cardiac CT image scale feature value to be calibrated is as follows: ； In the formula, is the scale feature value of the cardiac CT image to be calibrated, is the area covered by the heart front view of the heart CT image to be calibrated, is the area covered by the right heart view of the cardiac CT image to be calibrated, is the coverage area ratio of the left atrium and right atrium of the cardiac CT image to be calibrated, is the compensation factor for the area covered by the cardiac front view of the cardiac CT image to be calibrated, is the compensation factor for the area covered by the right cardiac view of the cardiac CT image to be calibrated, is the compensation factor for the ratio of the left atrium to the right atrium coverage area of ​​the cardiac CT image to be calibrated.

8. The method for geometric calibration of cardiac CT images based on machine learning according to claim 1, characterized in that: Based on the acquired cardiac CT image morphology data set to be calibrated, a comprehensive analysis is performed to obtain the morphology feature value of the cardiac CT image to be calibrated. The specific analysis process is as follows: Based on the acquired morphological data set of the cardiac CT image to be calibrated, a comprehensive analysis is performed to obtain the morphological feature value of the cardiac CT image to be calibrated, and the morphological feature value of the cardiac CT image to be calibrated is used as an analysis basis for determining whether the cardiac CT image to be calibrated needs to be geometrically calibrated.

9. A cardiac CT image geometric calibration system based on machine learning, characterized in that: A method for geometric calibration of cardiac CT images based on machine learning as described in any one of claims 1 to 8, comprising a geometric calibration defined feature range comparison module, a position feature value acquisition module, a scale feature value acquisition module, a morphology feature value acquisition module and a geometric calibration scheme comparison module, wherein: The geometric calibration defined feature range comparison module is used to obtain the cardiac CT image to be calibrated during atrial contraction, obtain a basic information data set of the object corresponding to the cardiac CT image to be calibrated, and compare and obtain the geometric calibration defined feature range of the cardiac CT image during contraction based on the obtained basic information data set of the object corresponding to the cardiac CT image to be calibrated; The position characteristic value acquisition module is used to acquire a position data set of the heart CT image to be calibrated, and obtain the position characteristic value of the heart CT image to be calibrated by comprehensive analysis based on the acquired position data set of the heart CT image to be calibrated; The scale characteristic value acquisition module is used to acquire the scale data set of the cardiac CT image to be calibrated, and obtain the scale characteristic value of the cardiac CT image to be calibrated by comprehensive analysis based on the acquired scale data set of the cardiac CT image to be calibrated; The morphological feature value acquisition module is used to acquire a morphological data set of the heart CT image to be calibrated, and based on the acquired morphological data set of the heart CT image to be calibrated, comprehensively analyze and obtain the morphological feature value of the heart CT image to be calibrated; The geometric calibration scheme comparison module is used to import the position characteristic values ​​of the cardiac CT image to be calibrated, the scale characteristic values ​​of the cardiac CT image to be calibrated and the morphological characteristic values ​​of the cardiac CT image to be calibrated into the cardiac CT image geometric calibration model, combine the characteristic range defined by the geometric calibration of the systolic cardiac CT image, and compare the geometric calibration scheme of the cardiac CT image that needs to be geometrically calibrated.

Images