A method, system and apparatus for calculating aortic cross-sectional area and dimensions

Abstract

The application discloses a kind of method, system and device for calculating aortic cross section axial and size, belong to medical detection technical field.The method for calculating aortic cross section axial and size is by reading aortic cross section original drawing and marking chart, picture information is data, then according to PCA algorithm principle calculation aortic cross section axial direction, finally linear interpolation is carried out to axial line, solve the problem of striding pixel point caused by traversing edge coordinate when pixel interval is larger, calculate the size of aortic cross section axial line.The application considers aortic cross section global pixel point simultaneously, constructs covariance matrix using the energy distribution of pixel point coordinates in each direction in two-dimensional space, and uses covariance matrix and stretches, rotates covariance matrix to decompose feature vector to determine the axial direction and size of aorta, realize the conversion from high dimension to low dimension of data to achieve the accurate calculation of aortic cross section axial direction and size.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of medical detection, and particularly relates to a method, system and device for calculating the axial direction and size of an aortic cross section. BACKGROUND

[0002] Aortic diseases are common diseases in cardiovascular diseases, such as aortic dissection, interstitial hematoma, atherosclerosis, aneurysm, Marfan syndrome, etc., which have high morbidity, great difficulty in treatment, and high mortality. In clinical practice, the axial direction and size of the aortic cross section based on medical imaging examination can be used as an important indicator for judging aortic diseases and selecting surgical plans, so the accurate calculation of the axial direction and size of the aortic cross section becomes particularly important.

[0003] In existing research, the calculation method of the axial direction and size of the aortic cross section is usually based on manual measurement of the sagittal and coronal positions of medical images, or the axial direction and size information of the aorta is determined according to the maximum and minimum Euclidean distances of the pixel points labeled on the cross section. The manual measurement method has multiple deviations due to the subjectivity of the measurer and the process of manual selection, and moreover, the method of directly determining the axial direction and size of the aorta based on the Euclidean distance between local pixel points without global consideration is too single in basis. These methods will have greater errors and disputes in the measurement of abnormal aortic conditions. Therefore, a method, system and device for calculating the axial direction and size of the aortic cross section are proposed to solve the above problems. SUMMARY

[0004] The present application aims to provide a method, system and device for calculating the axial direction and size of the aortic cross section, which aims to solve the problems of large deviation and low efficiency of the existing calculation method of the axial direction and size of the aortic cross section in the background art. To achieve the purpose, the technical solution adopted by the present application is as follows: a method for calculating the axial direction and size of the aortic cross section, which adopts the PCA algorithm principle to realize the dimension reduction function of high dimension, and can maximize the preservation of all information of the original variables in the dimension reduction process. The specific steps are as follows:

[0005] Step 1: read the aortic cross section original drawing and the labeled drawing, and dataize the picture information;

[0006] Step 2: calculate the axial direction of the aortic cross section according to the PCA algorithm principle, specifically:

[0007] S21: obtain all the horizontal and vertical coordinates of the labeled pixel points of the aortic cross section,

[0008] [X,Y]=[(x1,y1),(x2,y2),…,(x n ,y n )], where X=(x1,x2…x n), Y = (y1, y2, … y n ).

[0009] The labeled covariance matrix Mcov is calculated according to the following formula:

[0010]

[0011] where n is the number of pixels, X mean and Y mean are the horizontal and vertical coordinates after centering. The covariance matrix describes the degree of change between the horizontal and vertical coordinates of the label, and simultaneously represents the energy distribution of the horizontal and vertical coordinates of the label in the two-dimensional space of the original picture.

[0012] S22: Using the dimension reduction idea of the PCA algorithm, the aortic label point [X, Y] is mapped to a new two-dimensional space. The direction in the new two-dimensional space is the direction of the transformation matrix, and the energy of the labeled coordinates is concentrated in the transformed direction. Since matrix multiplication is equivalent to stretching or rotating the matrix, the stretched and rotated matrix is obtained by performing stretching and rotating operations on the covariance matrix Mcov:

[0013]

[0014] where V = (v1, v2) is the rotation matrix, which is also the eigenvector of the covariance matrix, S is the stretching matrix, and L obtained by formula reasoning is the eigenvalue diagonal matrix. Thus, the eigenvalues λ1 and λ2 inside the diagonal matrix and the corresponding eigenvectors v1 = [v 11 , v 12 ] and v2 = [v 21 , v 22 ] are obtained. Since v1 and v2 are orthogonal, the correlation and redundant information between the coefficients are reduced, the amount of information represented by each coefficient is maximized, the data conversion from high dimension to low dimension is achieved, and the precise calculation using the global pixel information of the aorta is achieved.

[0015] S23: The actual position of the principal axis of the aortic cross section is calculated according to the coordinate points of the labeled coordinates after centering, that is, the first principal axis:

[0016]

[0017] The second principal axis:

[0018]

[0019] Step 3: Calculate the aortic cross-sectional axis size, perform linear interpolation on the axis, and solve the problem of crossing pixels caused by traversing the edge coordinates when the pixel interval is large. The formula is as follows:

[0020] x_interp = interp[X min -1,X max +1,n×gap]

[0021]

[0022] x_interp and y_interp are the horizontal and vertical coordinates of the first or second principal axis after interpolation. The axes are iterated sequentially from both ends to check if they intersect with the aortic annotation. If they intersect, the boundary points of the aortic axis length are marked, resulting in the annotated pixel points [(X c1 ,Y c1 ),(X c2 ,Y c2 Therefore, the axial dimension of the aortic cross-section is:

[0023]

[0024] Spacing refers to the pixel spacing of pixels in the cross-section of the aorta.

[0025] Further describing the aforementioned scheme, in step S21, n is the number of pixels, and Xmean and Ymean are the centered horizontal and vertical coordinates, respectively. The algorithms for the positions and dimensions of the first and second principal axes of the aortic cross-section are the same.

[0026] According to one aspect of the present invention, a system for calculating the axial direction and dimensions of an aortic cross-section is provided, comprising an image input module, a storage module, an analysis and calculation module, and a display module. The image input module is used to input the original image and labeled image of the aortic cross-section, convert the image information into data and save it to the storage module. The analysis and calculation module reads the data in the storage module and counts the axial direction and dimensions of the aortic cross-section using the PCA algorithm, linear interpolation, and traversal algorithm. The display module is used to display the counting process and results of the analysis and calculation module.

[0027] Furthermore, the system's storage module stores the aortic cross-sectional axial and dimensional information of cases including aortic dissection, intramural hematoma, atherosclerosis, aneurysm, and Marfan syndrome. The analysis and calculation module compares the calculation results with the information on the characteristics of aortic dissection, intramural hematoma, atherosclerosis, aneurysm, and Marfan syndrome in the storage module and provides corresponding diagnostic results.

[0028] According to another aspect of the present invention, an apparatus for calculating the axial direction and dimensions of an aortic cross-section is also provided. The apparatus includes a camera, a processor, a memory, and a display screen. The camera, memory, and display screen are all electrically connected to the processor. The camera is used to record and store the original and annotated images of the aortic cross-section into the memory. The processor uses the aforementioned algorithm to calculate the axial direction and dimensions of the aortic cross-section from the data in the memory, and displays the technical process and results on the display screen.

[0029] Furthermore, the device also includes a communication module. The memory contains aortic cross-sectional axial and dimensional information for aortic dissection, intramural hematoma, atherosclerosis, aneurysm, and Marfan syndrome. The processor compares the technical results with the aortic cross-sectional axial and dimensional information for aortic dissection, intramural hematoma, atherosclerosis, aneurysm, and Marfan syndrome stored in the memory and uploaded to the network via the communication module, and provides corresponding diagnostic results. After the processor's diagnostic results are confirmed by the doctor, the corresponding aortic cross-sectional axial and dimensional information is stored in the memory and uploaded to the cloud server via the communication module, enabling iterative database upgrades and improving diagnostic accuracy.

[0030] Compared with existing technologies, this invention considers all pixels in the aortic cross-section simultaneously, constructs a covariance matrix using the energy distribution of pixel coordinates in various directions in two-dimensional space, and uses the covariance matrix and stretching and rotating it to decompose eigenvectors to determine the aortic axis and size. This achieves the transformation of data from high-dimensional to low-dimensional, thereby achieving accurate calculation of the aortic cross-section axis and size. Attached Figure Description

[0031] Figure 1 A schematic diagram illustrating a method for calculating the axial direction and dimensions of an aortic cross-section, provided by an embodiment of the present invention;

[0032] Figure 2 The original and annotated cross-sectional view of the aorta provided in the embodiments of the present invention;

[0033] Figure 3 A visual schematic diagram of the first spindle and the second spindle provided for an embodiment of the present invention;

[0034] Figure 4 This is a schematic diagram of the axis boundary points provided in an embodiment of the present invention.

[0035] The following are the labeling elements in the figure:

[0036] 1. Aortic cross-section; 2. First principal axis; 3. Second principal axis; 4. Boundary point. Detailed Implementation

[0037] To facilitate understanding of the present invention, it will now be described more fully with reference to the accompanying drawings. Preferred embodiments of the invention are shown in the drawings. However, the invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to provide a thorough and complete understanding of the disclosure of the invention.

[0038] The technical solution of this patent will be further described in detail below with reference to specific embodiments.

[0039] Example 1

[0040] like Figures 1-4 As shown, the present invention provides a method for calculating the axial direction and dimensions of the aortic cross-section, the specific steps of which are as follows:

[0041] Step 1: As Figure 2 As shown, the original and annotated images of the aortic cross-section 1 are read, and the image information is digitized.

[0042] Step 2: Calculate the axial direction of the aortic cross-section 1 according to the PCA algorithm principle, specifically:

[0043] First, obtain all the horizontal and vertical coordinates of the pixel marked on the cross-section of the aorta.

[0044] [X,Y]=[(x1,y1),(x2,y2),…,(x n ,y n )], where X=(x1,x2…x n ),Y=(y1,y2,…y n ).

[0045] The covariance matrix Mcov of the annotation is calculated according to the following formula:

[0046]

[0047] Where n is the number of pixels, and Xmean and Ymean are the centered x and y coordinates, respectively. The covariance matrix describes the degree of variation between the labeled x and y coordinates, and also represents the energy distribution of the labeled x and y coordinates in various directions in the two-dimensional space of the original image. This invention reduces two-dimensional image information to one-dimensional axes, using the PCA algorithm to achieve high-dimensionality reduction, and retains as much information as possible about the original variables during the reduction process.

[0048] S22: Using the dimensionality reduction concept of the PCA algorithm, the aortic annotation points [X,Y] are mapped to a new two-dimensional space. The direction in the new two-dimensional space is the direction of the transformation matrix, and the energy of the labeled coordinates is concentrated in the transformed direction. Since matrix multiplication is equivalent to stretching or rotating the matrix, stretching and rotating the covariance matrix Mcov yields the transformed matrix:

[0049]

[0050] Where V = (v1, v2) is the rotation matrix and also the eigenvector of the covariance matrix, S is the stretching matrix, and L, obtained through formula derivation, is the eigenvalue diagonal matrix. Thus, we obtain the eigenvalues ​​λ1 and λ2 inside the diagonal matrix and the corresponding eigenvector v1 = [v...]. 11 v 12 ] and v2 = [v 21 v 22 Because v1 and v2 are orthogonal, the correlation and redundant information between the coefficients are reduced, maximizing the amount of information represented by each coefficient. This achieves the transformation of data from high-dimensional to low-dimensional, enabling precise calculations using global pixel information of the aorta.

[0051] S23: Calculate the actual position of the main axis of the aortic cross-section 1, i.e., the first main axis 2, based on the coordinates after centering the labeled coordinates.

[0052]

[0053] Second spindle position 3:

[0054]

[0055] Step 3: Calculate the axis dimension of the aortic cross-section 1, and perform linear interpolation on the axis to solve the problem of crossing pixels caused by traversing edge coordinates when the pixel interval is large. The formula is as follows:

[0056] x_interp = interp[X min -1,X max +1, n×gap]

[0057]

[0058] x_interp and y_interp are the interpolated first principal axis 2 or second principal axis 3 horizontal and vertical coordinates. The axes are iterated sequentially from both ends to check if they intersect with the aortic annotation. If they intersect, the boundary point 4 of the aortic axis length is marked, resulting in the annotated pixel points [(X c1 ,Y c1 ),(X c2 ,Y c2)],like Figure 4 As shown, the axial dimension of the aortic cross-section 1 is:

[0059]

[0060] Spacing refers to the pixel interval of the aortic cross-section 1. The positions and dimensions of the first principal axis 2 and the second principal axis 3 of the aortic cross-section 1 are calculated using the same algorithm. Through the above steps, the axial direction and dimensions of the aortic cross-section 1 can be calculated automatically.

[0061] Example 2

[0062] This invention also provides a system for calculating the axial direction and dimensions of an aortic cross-section, comprising an image input module, a storage module, an analysis and calculation module, and a display module. The image input module inputs the original and annotated images of the aortic cross-section, converts the image information into data, and saves it to the storage module. The analysis and calculation module reads the data from the storage module and counts the axial direction and dimensions of the aortic cross-section using PCA algorithm principles, linear interpolation, and traversal algorithms. The display module displays the counting process and results from the analysis and calculation module. The system's storage module stores the axial direction and dimensions of the aortic cross-section, including cases of aortic dissection, intramural hematoma, atherosclerosis, aneurysm, and Marfan syndrome. The analysis and calculation module compares the calculation results with the information on the characteristics of aortic dissection, intramural hematoma, atherosclerosis, aneurysm, and Marfan syndrome in the storage module and provides corresponding diagnostic results. In addition to automatically calculating the axial direction and dimensions of the aortic cross-section, this system can also match corresponding cases based on the calculation results, reducing the workload of doctors.

[0063] Example 3

[0064] This invention also provides a device for calculating the axial direction and dimensions of an aortic cross-section. The device includes a camera, a processor, a memory, and a display screen. The camera, memory, and display screen are all electrically connected to the processor. The camera is used to input and store the original and annotated images of the aortic cross-section 1 into the memory. The processor calculates the axial direction and dimensions of the aortic cross-section 1 using the algorithm steps described in Example 1, and displays the process and results on the display screen. The device also includes a communication module. The memory contains information on the axial direction and dimensions of the aortic cross-section 1 from cases such as aortic dissection, intramural hematoma, atherosclerosis, aneurysm, and Marfan syndrome. The processor compares the results with the information on the axial direction and dimensions of the aortic cross-section 1 from cases such as aortic dissection, intramural hematoma, atherosclerosis, aneurysm, and Marfan syndrome from the memory and from the network via the communication module, and provides corresponding diagnostic results. After the processor's diagnostic results are confirmed by a doctor, the corresponding aortic cross-section axial direction and dimensions are stored in the memory and uploaded to a cloud server via the communication module, enabling database iteration and upgrades, and improving diagnostic accuracy. It should be noted that this device can be integrated into equipment such as CT scanners.

[0065] In summary, this invention simultaneously considers all pixels in the aortic cross-section, constructs a covariance matrix using the energy distribution of pixel coordinates in various directions in two-dimensional space, and uses the covariance matrix, along with stretching and rotating it, to decompose eigenvectors to determine the aortic axis and dimensions. This achieves data transformation from high-dimensional to low-dimensional, enabling accurate calculation of the aortic cross-section axis and dimensions, and can match approximate cases, greatly reducing the burden on doctors.

[0066] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0067] The above embodiments are only used to illustrate the present invention and are not intended to limit the present invention. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, all equivalent technical solutions also fall within the scope of the present invention, and the patent protection scope of the present invention should be defined by the claims.

Claims

1. A method for calculating the axial direction and dimensions of an aortic cross-section, characterized in that: Includes the following steps: Step 1: Read the original and annotated images of the aortic cross-section (1) and digitize the image information; Step 2: Calculate the axial direction of the aortic cross-section based on the PCA algorithm principle, specifically: S21: Obtain all horizontal and vertical coordinates of the labeled pixels in the aortic cross-section (1), and calculate the labeled covariance matrix Mcov according to the following formula: Where X is the set of abscissas of labeled pixels, X = (x1, x2, ..., xn), Y is the set of ordinates of labeled pixels, Y = (y1, y2, ..., yn), and n is the total number of labeled pixels; where Xmean is the average value of the abscissas of all labeled pixels after centering, and Ymean is the average value of the ordinates of all labeled pixels after centering. S22: Map the aortic annotation points [X,Y] to a new two-dimensional space, and perform stretching and rotation operations on the covariance matrix Mcov to obtain the transformed matrix: Where S is the stretching matrix, L is the eigenvalue diagonal matrix, λ1 and λ2 are the eigenvalues ​​inside the diagonal matrix, and V is the rotation matrix; S23: Calculate the actual position of the principal axis of the aortic cross section (1) based on the coordinate points after centering the labeled coordinates, i.e., the first principal axis: Second main axis: Step 3: Calculate the axial dimensions of the aortic cross-section (1), and perform linear interpolation on the axial dimensions to solve the problem of crossing pixels caused by traversing edge coordinates when the pixel interval is large. The formula is as follows: x_interp=interp[Xmin-1, Xmax+1, n×gap] y_interp=y(x_interp), Where x_interp is the set of interpolated principal axis abscissas and y_interp is the set of interpolated principal axis ordinates. The axis is iterated from both ends to check if it intersects with the aortic annotation. If it intersects, it is marked as the boundary point of the aortic axis length (4), and the annotation pixel points [(X c1 ,Y c1 ),(X c2 ,Y c2 Therefore, the axial dimension of the aortic cross-section is: Among them, spacing is the pixel spacing of the pixels in the cross-section of the aorta (1).

2. The method for calculating the axial direction and dimensions of the aortic cross-section according to claim 1, characterized in that: The positions and dimensions of the first main axis (2) and the second main axis (3) of the aortic cross section (1) are calculated in the same way.

3. A system for calculating the axial direction and dimensions of the aortic cross-section, characterized in that: It includes an image input module, a storage module, an analysis and calculation module, and a display module. The image input module is used to input the original image and labeled image of the aortic cross section, convert the image information into data and save it to the storage module. The analysis and calculation module reads the data in the storage module and calculates the axial direction and dimensions of the aortic cross section (1) using the algorithm steps as described in claim 1. The display module is used to display the calculation process and results of the analysis and calculation module.

4. The system for calculating the axial direction and dimensions of the aortic cross-section according to claim 3, characterized in that: The storage module stores the axial direction and dimensions of the aortic cross-section, including aortic dissection, intramural hematoma, atherosclerosis, aneurysm, and Marfan syndrome. The analysis and calculation module compares the calculation results with the characteristic information of aortic dissection, intramural hematoma, atherosclerosis, aneurysm, and Marfan syndrome in the storage module and provides corresponding diagnostic results.

5. A device for calculating the axial direction and dimensions of the aortic cross-section, characterized in that: It includes a camera, a processor, a memory, and a display screen. The camera, memory, and display screen are all electrically connected to the processor. The camera is used to input and store the original and annotated images of the aortic cross-section (1) into the memory. The processor uses the algorithm steps as described in claim 1 to calculate the axial direction and dimensions of the aortic cross-section (1) based on the data in the memory, and displays the calculation process and results on the display screen.

6. The apparatus for calculating the axial direction and dimensions of the aortic cross-section according to claim 5, characterized in that: It also includes a communication module. The memory contains axial and dimensional information of the aortic cross section (1) for aortic dissection, intramural hematoma, atherosclerosis, aneurysm, and Marfan syndrome. The processor compares the calculation results with the axial and dimensional information of the aortic cross section (1) for aortic dissection, intramural hematoma, atherosclerosis, aneurysm, and Marfan syndrome in the memory and on the network through the communication module, and gives the corresponding diagnostic results.

7. The apparatus for calculating the axial direction and dimensions of the aortic cross-section according to claim 6, characterized in that: After the results of the processor diagnosis are confirmed by the doctor, the corresponding aortic cross-section (1) axial and dimensional information is stored in the memory and uploaded to the cloud server through the communication module.

Images