Cervical vertebra nuclear magnetic resonance image preprocessing method and system

Abstract

The invention relates to the technical field of nuclear magnetic image processing, and discloses a cervical vertebra nuclear magnetic resonance image preprocessing method and system, and the method comprises the steps: obtaining a cervical vertebra magnetic resonance image, and carrying out the preprocessing of the cervical vertebra magnetic resonance image; performing primary segmentation on the preprocessed image, and extracting a spinal cord region; processing the spinal cord area, and judging whether an abnormal position exists or not; if no abnormal position exists, carrying out secondary segmentation on the image according to a first strategy; and if there is an abnormal position, carrying out secondary segmentation on the image according to a second strategy. The method comprises the following steps: firstly, identifying a spinal cord region which is easy to process, processing the boundary of the spinal cord region to judge whether an abnormal position exists, performing directional segmentation on the possible abnormal position, and processing to make image details clear; therefore, the personnel can know the detailed situation of the abnormal position at the first time while checking the cervical vertebra nuclear magnetic image, and the personnel can be assisted to quickly judge.

Description

Technical Field

[0001] The present invention relates to the technical field of nuclear magnetic resonance image processing, and in particular to a method and system for preprocessing cervical nuclear magnetic resonance images. Background Art

[0002] Cervical disc herniation, spinal stenosis, degenerative lesions, etc. are common cervical diseases. MRI (magnetic resonance imaging) is a direct imaging examination method that analyzes nuclear magnetic resonance images to obtain specific lesions. Due to large individual differences, the width of each person's spinal canal is inconsistent. The existing method of simply extracting spinal canal width data to judge lesions is inaccurate. The original image is used for identification and extraction. The processing model is large, the operation time is long, and targeted extraction cannot be performed. Overall, the processing of nuclear magnetic resonance images of the cervical spine is time-consuming and labor-intensive, and the effect is poor. Summary of the invention

[0003] In view of the deficiencies in the prior art, the object of the present invention is to provide a method and system for preprocessing cervical magnetic resonance images.

[0004] In order to achieve the above object, the present invention provides the following technical solutions:

[0005] A method for preprocessing cervical vertebrae magnetic resonance images, comprising the steps of:

[0006] Acquire a cervical vertebra magnetic resonance image, and preprocess the cervical vertebra magnetic resonance image;

[0007] The preprocessed image is segmented at the first level to extract the spinal cord area;

[0008] Treat the spinal cord area to determine if there is any abnormal location;

[0009] If there is no abnormal location, the image is segmented at the second level according to the first strategy;

[0010] If there is an abnormal position, the image is segmented at the second level according to the second strategy.

[0011] In the present invention, preferably, the first-level segmentation includes:

[0012] Several connected domains are determined through the recognition algorithm, and then judged according to the range and position of each connected domain. The connected domain range runs vertically through the entire image, and the mark located in the middle of the image is the spinal cord area.

[0013] In the present invention, preferably, the spinal cord region is subjected to boundary fitting, and it is determined whether there is abnormal data according to the change of the line segment after boundary fitting. If there is abnormal data, the corresponding spinal cord region is first marked, otherwise it is not marked.

[0014] In the present invention, preferably, if there is a first mark, the image is segmented at the second level according to a first strategy, wherein the first strategy includes:

[0015] Segment the cervical spine magnetic resonance image laterally based on the first marker expansion method;

[0016] Image enhancement and abnormal location annotation are performed on the segmented cervical spine magnetic resonance images.

[0017] In the present invention, preferably, the boundary fitting uses the least square method to fit the boundary of the spinal cord area into a plurality of straight lines, and determines whether there is an abnormal position based on the change in slope between adjacent straight lines.

[0018] In the present invention, preferably, the boundary fitting is based on the pixel coordinates of the spinal cord region in the original cervical magnetic resonance image, and a number of pixel coordinates on the boundary of the spinal cord region are selected at set intervals, and a number of boundary lines are obtained by least squares fitting.

[0019] In the present invention, preferably, the first marker expansion type transverse segmentation of the cervical vertebrae magnetic resonance image is based on the first marker as the center line, and a number of pixels are expanded vertically and horizontally to determine the segmentation area, and then segmentation is performed to segment out the corresponding abnormal position.

[0020] In the present invention, preferably, it is determined whether the spinal cord region has a first marker. If there is no first marker, a second strategy is used for secondary segmentation. The second strategy includes:

[0021] The left side of the spinal cord region is selected, and the vertebral body specific area is determined for the left side region;

[0022] The vertebral bodies and intervertebral discs are identified in specific areas of the vertebral body, and each vertebral body and intervertebral disc is numbered.

[0023] In the present invention, preferably, the numbering is performed in sequence from top to bottom based on the pixel positions corresponding to each vertebral body and intervertebral disc.

[0024] The cervical vertebrae MRI image preprocessing system comprises a preprocessing unit, a first processing unit, a judgment unit, a second processing unit, a third processing unit and a display unit. The preprocessing unit is used to preprocess the cervical vertebrae MRI image. The first processing unit is used to perform primary segmentation on the preprocessed cervical vertebrae MRI image and extract the spinal cord area. The judgment unit is used to process the spinal cord area and judge whether there is an abnormal position. If there is an abnormality, the second processing unit performs secondary segmentation on the image according to the first strategy. If there is no abnormality, the third unit performs secondary segmentation on the image according to the second strategy. The display unit is used to display the segmented image.

[0025] Compared with the prior art, the present invention has the following beneficial effects:

[0026] The method of the present invention first identifies the spinal cord area that is easy to process, processes the boundary of the spinal cord area to determine whether there is an abnormal position, performs direction segmentation on the possible abnormal position, and makes the image details clear through processing, so that personnel can immediately understand the details of the abnormal position while viewing the cervical spine magnetic resonance image. Assisted personnel can quickly and better grade cervical spondylosis, evaluate the intervertebral disc, intervertebral foramen, and spinal cord compression, and further segment the vertebral body and intervertebral disc when there is no abnormality in the spinal cord area to display the decomposed detail map, so that personnel can further understand the situation through the detail map at any time when viewing the overall image, without the need for personnel to manually enlarge the map, and at the same time, through specific segmentation enhancement, the slow processing speed of the existing direct recognition model is avoided, the hardware conditions are small, and the feasibility is strong. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 The figure is a flow chart of a cervical spine magnetic resonance image preprocessing method according to the present invention.

[0028] Figure 2 This is a schematic diagram of image segmentation using the first strategy described in the present invention.

[0029] In the figure, 1 is the original cervical spine MRI image, 2 is the extracted spinal cord area, and 3 is the segmented abnormal position. DETAILED DESCRIPTION

[0030] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0031] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used herein in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more related listed items.

[0032] Please also see Figure 1 and Figure 2A preferred embodiment of the present invention provides a cervical vertebrae MRI image preprocessing method, which mainly targets cervical vertebrae MRI images, first extracts the spinal cord region that is easy to identify and extract, and determines whether there is an abnormal position by processing the spinal cord region. Different image segmentation processes are performed based on different judgments to meet different image requirements. The overall process is simple and fast, ensuring the quality of the processed image. At the same time, the targeted image processing speed is fast and the applicability to the processing hardware is strong. Specifically, the steps are as follows:

[0033] Acquire a cervical vertebra magnetic resonance image, and pre-process the cervical vertebra magnetic resonance image;

[0034] The preprocessed image is segmented at the first level to extract the spinal cord area;

[0035] Treat the spinal cord area to determine if there is any abnormal location;

[0036] If there is no abnormal location, the image is segmented at the second level according to the first strategy;

[0037] If there is an abnormal position, the image is segmented at the second level according to the second strategy.

[0038] Specifically, a cervical vertebra magnetic resonance image is first obtained, and a bilateral filtering algorithm is used to preprocess the cervical vertebra resonance image to remove noise from the image and enhance the edges and details in the image.

[0039] Then, the preprocessed image is segmented at the first level, and several connected domains are determined through the recognition algorithm. The connected domains are judged according to their range and position, and the connected domains that run through the entire image vertically and are located in the middle of the image are marked as the spinal cord area.

[0040] In this embodiment, the watershed algorithm is used to segment the preprocessed cervical magnetic resonance image, and the image is divided into several regions. In order to avoid over-segmentation of the watershed algorithm, the segmented image is subjected to Gaussian smoothing. Then, according to the position and range of each region in the image, the range of each connected domain is judged, and the connected domain range that runs through the entire image vertically, has a width less than the preset range, and is located in the middle of the image horizontally is marked as the spinal cord region.

[0041] Then the spinal cord region is processed, the boundary of the spinal cord region is fitted, and whether there is abnormal data is determined based on the change of the line segment after the boundary fitting. If there is abnormal data, the corresponding spinal cord region is marked for the first time, and the first mark is a horizontal straight line, otherwise it is not marked.

[0042] In this embodiment, the boundary fitting is based on the pixel coordinates of the spinal cord region in the original cervical vertebrae magnetic resonance image. A number of pixel coordinates on the boundary of the spinal cord region are selected at set intervals, and a number of boundary lines are obtained by least squares fitting. The straight lines are represented as:

[0043] y=a+bx

[0044] Where (x, y) is the coordinate of one pixel point, a and b are parameters, and a and b in each straight line are determined by the least squares method. According to the slope of the current straight line, the pixel point P of the spinal cord boundary corresponding to the straight line can be known. i The corresponding direction is By judging the directions between several adjacent pixels on the spinal cord boundary, when the corresponding directions are inconsistent, the position of the pixel is marked, and a horizontal straight line is drawn through the point. The straight line is the first mark. If the directions of the pixels are consistent, no marking is performed.

[0045] When there is a first mark, the image is segmented at the second level according to the first strategy. The first strategy includes: based on the first mark, the cervical spine magnetic resonance image is segmented horizontally in an extended manner, that is, the first mark is used as the center line, and a number of pixels are extended vertically and horizontally to determine the segmentation area, and then segmentation is performed to segment the corresponding abnormal position. The specific number of extended pixels can be set in advance according to needs. The segmented cervical spine magnetic resonance image is enhanced and the abnormal position is marked. Through this method, the image can be processed quickly, and the abnormal position can be extracted for enlargement, enhancement and marking, so that personnel can observe the abnormal position at the first time and assist them in making a quick judgment.

[0046] When there is no first mark, the second strategy is used for secondary segmentation. The second strategy includes:

[0047] The left side of the spinal cord region is selected, and the vertebral body specific area is determined for the left side region;

[0048] Vertebrae and intervertebral discs are identified in specific areas of the vertebrae, and each vertebrae and intervertebral disc is numbered. The numbering is based on the pixel position corresponding to each vertebrae and intervertebral disc, and is numbered from top to bottom. The second strategy also enhances the image. Since the contrast of the MRI image is low and the vertebrae and intervertebral discs are relatively dark, filtering is first performed to smooth the image and eliminate interference; then normalization is performed to make the grayscale of the image evenly distributed, improve the contrast and layering, and increase the recognizability of details. After this processing, local detail images can be displayed around the cervical spine magnetic resonance image without manual zooming, avoiding the unclear manually zoomed images and making it convenient for personnel to view the detail images.

[0049] Another preferred embodiment of the present invention provides a cervical vertebra magnetic resonance image preprocessing system, including a preprocessing unit, a first processing unit, a judgment unit, a second processing unit, a third processing unit, and a display unit. The preprocessing unit is used to preprocess the cervical vertebra magnetic resonance image. The first processing unit is used to perform a primary segmentation on the preprocessed cervical vertebra magnetic resonance image to extract the spinal cord area. The judgment unit is used to process the spinal cord area to determine whether there is an abnormal position. If there is an abnormality, the second processing unit performs a secondary segmentation on the image according to the first strategy. If there is no abnormality, the third unit performs a secondary segmentation on the image according to the second strategy. The display unit is used to display the segmented image.

[0050] In some other preferred embodiments of the present invention, a computer-readable storage medium is provided, which stores a computer program. When the computer program is executed by a processor, the processor executes the steps of the method in the above embodiment.

[0051] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.

[0052] The above description is a detailed description of the preferred feasible embodiments of the present invention, but the embodiments are not intended to limit the scope of the patent application of the present invention. All equivalent changes or modified changes completed under the technical spirit suggested by the present invention should fall within the patent scope covered by the present invention.

Claims

1. A method for preprocessing cervical vertebrae magnetic resonance images, characterized in that: Includes steps: Acquire a cervical vertebra magnetic resonance image, and preprocess the cervical vertebra magnetic resonance image; The preprocessed image is segmented at the first level to extract the spinal cord area; Treat the spinal cord area to determine if there is any abnormal location; If there is no abnormal location, the image is segmented at the second level according to the first strategy; If there is an abnormal position, the image is segmented at the second level according to the second strategy.

2. The cervical vertebrae magnetic resonance image preprocessing method according to claim 1, characterized in that: The first-level segmentation includes: Several connected domains are determined through the recognition algorithm, and then judged according to the range and position of each connected domain. The connected domain range runs vertically through the entire image, and the mark in the middle of the image is the spinal cord area.

3. The cervical vertebra nuclear magnetic resonance image preprocessing method according to claim 2, characterized in that: Boundary fitting is performed on the spinal cord region, and whether there is abnormal data is determined according to the change of the line segment after boundary fitting. If there is abnormal data, the corresponding spinal cord region is first marked, otherwise it is not marked.

4. The cervical vertebrae magnetic resonance image preprocessing method according to claim 3, characterized in that: If there is a first mark, the image is segmented at the second level according to a first strategy, wherein the first strategy includes: Segment the cervical spine magnetic resonance image laterally based on the first marker expansion method; Image enhancement and abnormal location marking are performed on the segmented cervical spine magnetic resonance images.

5. The cervical vertebrae magnetic resonance image preprocessing method according to claim 4, characterized in that: The boundary fitting uses the least square method to fit the boundary of the spinal cord area into a plurality of straight lines, and determines whether there is an abnormal position according to the change in the slope between adjacent straight lines.

6. The cervical vertebra nuclear magnetic resonance image preprocessing method according to claim 5, characterized in that: The boundary fitting is based on the pixel coordinates of the spinal cord region in the original cervical vertebra magnetic resonance image, and a number of pixel coordinates on the boundary of the spinal cord region are selected at set intervals, and a number of boundary lines are obtained by least squares fitting.

7. The cervical vertebrae magnetic resonance image preprocessing method according to claim 3, characterized in that: The first marker expansion-type transverse segmentation of the cervical vertebrae magnetic resonance image is to use the first marker as the center line, expand a number of pixels vertically and horizontally to determine the segmentation area, and then perform segmentation to segment the corresponding abnormal position.

8. The cervical vertebrae magnetic resonance image preprocessing method according to claim 1, characterized in that: Determine whether the spinal cord region has a first mark. If there is no first mark, use a second strategy to perform secondary segmentation. The second strategy includes: The left side of the spinal cord region is selected, and the vertebral body specific area is determined for the left side region; The vertebral bodies and intervertebral discs are identified in specific areas of the vertebral body, and each vertebral body and intervertebral disc is numbered.

9. The cervical vertebrae magnetic resonance image preprocessing method according to claim 8, characterized in that: The numbering is performed from top to bottom based on the pixel positions corresponding to each vertebral body and intervertebral disc.

10. A system for implementing the cervical spine magnetic resonance image preprocessing method according to any one of claims 1 to 9, characterized in that: The invention comprises a preprocessing unit, a first processing unit, a judgment unit, a second processing unit, a third processing unit and a display unit. The preprocessing unit is used to preprocess the cervical vertebrae magnetic resonance image. The first processing unit is used to perform primary segmentation on the preprocessed cervical vertebrae magnetic resonance image and extract the spinal cord area. The judgment unit is used to process the spinal cord area and judge whether there is an abnormal position. If there is an abnormality, the second processing unit performs secondary segmentation on the image according to the first strategy. If there is no abnormality, the third unit performs secondary segmentation on the image according to the second strategy. The display unit is used to display the segmented image.

Images