A three-dimensional cardiac positioning method based on MR transverse axial image segmentation

Through the three-dimensional positioning method based on MR horizontal axis image segmentation, the ‘region growth segmentation algorithm’ and positioning surface parameter adjustment are used to solve the problem of complexity and low accuracy of manual positioning, and efficient and accurate heart positioning is achieved.

CN114419071BActive Publication Date: 2024-11-08SOUTHEAST UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210075808.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-01-23
Publication Date
2024-11-08
Estimated Expiration
2042-01-23

AI Technical Summary

Technical Problem

The existing manual cardiac positioning methods are complex and rely on the experience of doctors, especially in the "two-chakra->four-chakra" step, which is prone to errors, affecting the accuracy of positioning and process efficiency.

Method used

Using a three-dimensional positioning method based on MR horizontal axis image segmentation, semi-automatic heart region segmentation is used to reconstruct the three-dimensional heart image, and the scanning surface parameters are calculated by adjusting the positioning surface parameters to simplify the long axis and short axis of the heart. , four-chamber center and other scanning surfaces are positioned.

Benefits of technology

It improves the accuracy and efficiency of cardiac positioning, shortens the positioning process time, and stacks the three-dimensional cardiac images through multiple MR horizontal axis images to enable doctors to more intuitively determine the heart morphology and scanning surface position.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114419071B_ABST
    Figure CN114419071B_ABST
Patent Text Reader

Abstract

The present invention discloses a three-dimensional heart positioning method based on MR transverse axial image segmentation, which scans multi-layer MR transverse axial images, uses a "region growing segmentation algorithm" to segment the heart region, reconstructs a three-dimensional heart image, manually adjusts the positioning surface, calculates the scanning surface parameters, and inputs them into the scanning instrument to complete the scanning of the heart's four-chamber heart, transverse axis, and long axis. Compared with the manual positioning method, the present invention only needs to scan the transverse axial chest image, obtain the scanning surface parameters after post-processing, and input them into the magnetic resonance scanning system, so as to directly scan the long axis, short axis, four-chamber heart and other scanning surfaces of the heart, shorten the time consumed by the positioning process, and improve the accuracy of heart positioning.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to a three-dimensional heart positioning method based on MR transverse axial image segmentation, and belongs to the field of computer image processing. Background Art

[0002] Magnetic resonance imaging is an important means of diagnosing heart disease, and it combines the advantages of other detection methods (such as CT and ultrasound). For example, the general morphology of the heart can be displayed through conventional sequences; myocardial function can be evaluated through movie sequences; and myocardial diseases can be diagnosed through first-pass perfusion and delayed enhancement. However, there is a necessary step before these examinations - cardiac positioning.

[0003] When analyzing anatomical structures such as the brain and liver, clinicians usually use three directions, namely transverse, sagittal and coronal, for tomography. However, this method cannot be used for cardiac positioning scans, because the heart is tilted relative to the vertical direction of the human body, and the cardiac tomographic images scanned using the transverse, coronal and sagittal directions are difficult to use as a basis for diagnosis. Accordingly, the heart needs to be scanned in three important directions, including the long axis, short axis and four-chamber heart. The process of finding these three directions is called cardiac positioning.

[0004] At present, the main method of cardiac positioning is manual positioning, that is, drawing positioning lines on the two-dimensional positioning image, and positioning step by step according to some step specifications. This method has a relatively complicated process, and the general steps are "scanning transverse axis -> two-chamber heart -> four-chamber heart -> short axis -> long axis". The positioning accuracy of each step will affect the subsequent process, especially the "two-chamber heart -> four-chamber heart" step is prone to errors, so the manual positioning method is difficult to use and depends on the doctor's experience. Summary of the invention

[0005] In view of the problems existing in the manual positioning method, the present invention proposes a three-dimensional heart positioning method based on MR transverse axial image segmentation, providing doctors with another solution for heart positioning.

[0006] The present invention adopts the following technical solutions to solve the above problems:

[0007] The present invention provides a three-dimensional cardiac positioning method based on MR transverse axial image segmentation, and the specific steps are as follows:

[0008] Step 1: Scan the MR transverse axial plane and select the appropriate scanning slice thickness and number of slices so that the transverse axial plane covers the entire heart;

[0009] Step 2: Use the “region growing segmentation algorithm” to segment the heart region of each image. The “region growing segmentation algorithm” is a semi-automatic segmentation method. The process is as follows: Figure 1As shown, take the first pixel from the queue, use this pixel as the "seed", search for pixels around it, if the pixel tag = 0, and the pixel value difference between the pixel tag and the seed is less than the "threshold", then add the pixel to the queue and set the tag to 1;

[0010] Step 3, reconstructing a three-dimensional heart image, adjusting the pixel interval of the heart mask obtained by the segmentation in step 2, and stacking them into a three-dimensional heart image; the specific details of reconstructing the three-dimensional heart image mentioned in step 3 are as follows: the software reads the PixelSpacing field in the DICOM file as the pixel interval of the x-axis and the y-axis, and uses the difference between the ImagePositionPatient fields in two adjacent DICOM files as the pixel interval of the z-axis, synthesizes the heart mask obtained in step 2 into a three-dimensional array, multiplies it by the pixel interval, and uses VTK in the software to render a three-dimensional heart image with actual size.

[0011] Step 4: Manually adjust the positioning surface and calculate the scanning surface parameters. The relationship between the positioning surface and the scanning surface is as follows: Figure 2 As shown in the figure, the positioning surface is a parallelogram that can be adjusted by the user, and the scanning surface is a square with a specified side length, centered on the positioning surface. The user can change the shape of the positioning surface by adjusting the coordinates of the starting point of the positioning surface, the direction vectors and lengths of the horizontal and vertical axes, and move the positioning surface to the specified position, set the side length of the scanning surface, and calculate the scanning surface parameters (upper left corner coordinate A, horizontal axis X direction vector Y-axis vector Plane Normal ). Known positioning surface horizontal axis M direction vector Length L m , vector in the direction of the vertical axis N Length L n , the coordinate of the upper left corner P and the side length of the scanning surface L. At the same time, it is specified that normalize() represents standardization (converting to unit vector), and the × operator represents cross product. Represents the normal vector of the positioning surface and the scanning surface. The derivation process of the scanning surface parameters is as follows:

[0012]

[0013]

[0014]

[0015]

[0016]

[0017]

[0018]

[0019]

[0020]

[0021]

[0022]

[0023] Step 5, guide the scanning surface parameters (including the upper left corner coordinates, the horizontal axis direction vector, the vertical axis direction vector, and the surface normal vector) into the magnetic resonance scanning system to perform nuclear magnetic resonance scanning.

[0024] Compared with the prior art, the present invention has the following advantages: (1) by using the "region growing algorithm" to perform semi-automatic segmentation of chest images, the efficiency of doctors in segmenting the heart area can be improved; (2) by stacking a plurality of MR transverse axial images to form a three-dimensional heart image, doctors can more intuitively observe the heart morphology and determine the position of the heart scanning plane; (3) by adjusting the parameters of the positioning plane, the positioning plane can be quickly adjusted to the desired position, and the parameters of the scanning plane can be calculated and input into the magnetic resonance scanning system for scanning; (4) compared with the manual positioning method, this method only needs to scan the transverse axial chest image, and after post-processing to obtain the scanning plane parameters, the long axis, short axis, four-chamber heart and other scanning planes of the heart can be directly scanned, thereby shortening the time consumed by the positioning process and improving the accuracy of heart positioning. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 It is the flow chart of the region growing segmentation algorithm;

[0026] Figure 2 It is a schematic diagram of the relationship between the positioning surface and the scanning surface;

[0027] Figure 3 It is a schematic diagram of the process of the present invention;

[0028] Figure 4 It is a schematic diagram of the overall positioning method of the present invention. DETAILED DESCRIPTION

[0029] The present invention is further explained below in conjunction with specific embodiments. It should be understood that these embodiments are only used to illustrate the present invention and are not used to limit the scope of the present invention. After reading the present invention, various equivalent forms of modifications to the present invention by those skilled in the art all fall within the scope defined by the claims attached to this application.

[0030] Embodiment 1: The present invention provides a method for three-dimensional positioning of the heart based on MR transverse axial image segmentation, and the specific steps are as follows:

[0031] Step 1: Scan the MR transverse axial plane and select the appropriate scanning slice thickness and number of slices so that the transverse axial plane covers the entire heart;

[0032] Step 2: Use the “region growing segmentation algorithm” to segment the heart region of each image. The “region growing segmentation algorithm” is a semi-automatic segmentation method. The process is as follows: Figure 1 As shown, take the first pixel from the queue, use this pixel as the "seed", search for pixels around it, if the pixel tag = 0, and the pixel value difference between it and the seed is less than the "threshold", then add the pixel to the queue and set the tag to 1;

[0033] Step 3, reconstructing a three-dimensional heart image, adjusting the pixel interval of the heart mask obtained by the segmentation in step 2, and stacking them into a three-dimensional heart image; the specific details of reconstructing the three-dimensional heart image mentioned in step 3 are as follows: the software reads the PixelSpacing field in the DICOM file as the pixel interval of the x-axis and the y-axis, and uses the difference between the ImagePositionPatient fields in two adjacent DICOM files as the pixel interval of the z-axis, synthesizes the heart mask obtained in step 2 into a three-dimensional array, multiplies it by the pixel interval, and uses VTK in the software to render a three-dimensional heart image with actual size.

[0034] Step 4: Manually adjust the positioning surface and calculate the scanning surface parameters. The relationship between the positioning surface and the scanning surface is as follows: Figure 2 As shown in the figure, the positioning surface is a parallelogram that can be adjusted by the user, and the scanning surface is a square with a specified side length, centered on the positioning surface. The user can move the positioning surface to a specified position, set the side length of the scanning surface, and calculate the scanning surface parameters (upper left corner coordinate A, horizontal axis X direction vector Y-axis vector Plane Normal ). Known positioning surface horizontal axis M direction vector Length L m , vector in the direction of the vertical axis N Length L n , the coordinate of the upper left corner P and the side length of the scanning surface L. At the same time, it is specified that normalize() represents standardization (converting to unit vector), and the × operator represents cross product. Represents the normal vector of the positioning surface and the scanning surface. The derivation process of the scanning surface parameters is as follows:

[0035]

[0036]

[0037]

[0038]

[0039]

[0040]

[0041]

[0042]

[0043]

[0044]

[0045]

[0046] Step 5, import the scanning surface parameters (including the upper left corner coordinates, the horizontal axis direction vector, the vertical axis direction vector, and the surface normal vector) into the instrument to perform nuclear magnetic resonance scanning.

[0047] The whole process diagram of the invention is as follows Figure 3 shown.

[0048] Specific embodiment: A three-dimensional cardiac positioning method based on MR transverse axial image segmentation, the specific steps are as follows:

[0049] Step 1: Scan the MR transverse axis. After the patient enters the scanner, the doctor sets the slice thickness to 8 mm, the number of slices to 12, and makes the positioning line cover the entire heart, scans, and saves the obtained image in DICOM format;

[0050] Step 2, segment the heart area of ​​the image. Import the DICOM file obtained in step 1 into the software, adjust the image window width and window position, and use the "region growing segmentation algorithm" for segmentation. The segmentation process is as follows: click a pixel in the image with the mouse, and the pixel is regarded as a "seed". Search around to find pixels with a pixel value difference less than the "threshold" and classify them as heart areas. If the segmentation is inaccurate, click Undo, adjust the "threshold", and continue the segmentation operation. Finally, 12 heart segmentation images are obtained;

[0051] Step 3: Reconstruct a three-dimensional cardiac image. The software reads the PixelSpacing field in the DICOM file as the pixel spacing of the x-axis and y-axis, uses the difference between the ImagePositionPatient fields in two adjacent DICOM files as the pixel spacing of the z-axis, synthesizes the cardiac mask obtained in step 2 into a three-dimensional array, multiplies it by the pixel spacing, and reconstructs a three-dimensional cardiac image with actual size in the software;

[0052] Step 4: Manually adjust the positioning surface and calculate the scanning surface parameters. In the software, adjust the starting point, horizontal axis, and vertical axis of the positioning surface so that the positioning surface is located at the four-chamber heart position of the heart. Then adjust the side length of the scanning surface, and obtain the scanning surface parameters (upper left corner coordinates, horizontal axis direction vector, vertical axis direction vector, plane normal vector) through the transformation relationship between the positioning surface and the scanning surface.

[0053] Step 5, import the scanning surface parameters into the scanner software and perform MRI scanning.

[0054] Effect evaluation:

[0055] The present invention proposes a three-dimensional heart positioning method based on MR transverse axial image segmentation, which provides an effective solution for heart positioning. The quality of the heart positioning map obtained by the present invention depends on the accuracy of the positioning surface.

[0056] The above description is only a specific implementation of the present invention, but the protection scope of the present invention is not limited thereto. Any person familiar with the technology can understand and think of any changes or substitutions within the technical scope disclosed by the present invention, which should be included in the scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A three-dimensional cardiac positioning method based on MR transverse axial image segmentation, characterized in that: The specific steps are as follows: Step 1: Scan the MR transverse axial plane and select the appropriate scanning slice thickness and number of slices so that the transverse axial plane covers the entire heart; Step 2: Use the "region growing segmentation algorithm" to segment the heart region of each image; Step 3, reconstructing a three-dimensional heart image, using the heart mask obtained by segmentation in step 2 to reconstruct a three-dimensional heart image; Step 4, manually adjust the positioning surface and calculate the scanning surface parameters; Step 5: Import the scanning surface parameters into the instrument and perform MRI scanning. The "region growing segmentation algorithm" mentioned in step 2 is written into a software. The software is written using the pyqt5 framework. The functions are as follows: after importing the image, adjust the window width and window position, click a pixel in the image with the mouse, and the pixel is regarded as a "seed". Search around to find pixels with a pixel value difference less than the "threshold" and classify them as the heart area. If the segmentation is inaccurate, click Undo, adjust the "threshold", and continue the segmentation operation; The scanning surface parameters mentioned in step 4 are calculated based on the positioning surface parameters. It is known that the scanning surface is a square, and its parameters include the coordinates A of the starting point, the direction vector of the horizontal axis The direction vector of the vertical axis The side length is L, and the positioning surface is a parallelogram. The parameters include the coordinates of the upper left corner P and the direction vector of the horizontal axis. Length L m , the direction vector of the vertical axis Length L n , and it is specified that normalize() means to standardize the vector, and the × operator means the cross product. Represents the normal vectors of the positioning surface and the scanning surface. The derivation process of the scanning surface parameters is as follows:

2. The method for three-dimensional cardiac positioning based on MR transverse axial image segmentation according to claim 1, characterized in that: The specific details of the reconstructed three-dimensional heart image mentioned in step 3 are as follows: the software reads the PixelSpacing field in the DICOM file as the pixel interval of the x-axis and y-axis, takes the difference between the ImagePositionPatient fields in two adjacent DICOM files as the pixel interval of the z-axis, synthesizes the heart mask obtained in step 2 into a three-dimensional array, multiplies it by the pixel interval, and uses VTK in the software to render a three-dimensional heart image with actual size.

3. The method for three-dimensional cardiac positioning based on MR transverse axial image segmentation according to claim 1, characterized in that: The manual adjustment of the positioning surface mentioned in step 4 is implemented using software. The implementation details are as follows: the positioning surface is a parallelogram. The parameters that determine its position and shape include the starting point (x, y, z coordinates), the horizontal and vertical axes, namely the angle with the x-axis, the angle with the y-axis, the angle with the z-axis, and the length. By changing these parameters, the shape of the positioning surface can be changed and it can be moved to the specified position.