Method and system for automatic bedplate removal in ct images

Abstract

This invention provides a method and system for automatically removing bed slabs from CT images, relating to the field of image segmentation technology. The method includes: acquiring the three-dimensional dimensions X, Y, and Z of the CT image; constructing a 2D image of size X and Y; traversing pixel values ​​along the z-axis of the CT image and assigning the maximum pixel value at (x, y) to (x, y) in the 2D image; performing threshold segmentation on the 2D image based on the CT values ​​to extract the bed slab and human tissue regions, obtaining a binary image; performing seed point region growing processing on the binary image to extract the human tissue regions, obtaining a mask_2D image; stitching together Z mask_2D images along the z-axis of the CT image to obtain a mask_3D image; and performing image template segmentation on the CT image based on the mask_3D image to obtain a 3D image after removing the bed slab. This invention achieves automatic and rapid removal of bed slabs from CT images through image segmentation technology.

Description

Technical Field

[0001] This invention belongs to the field of image segmentation technology, and in particular relates to a method and system for automatically removing bed boards from CT images. Background Technology

[0002] During a CT scan, the patient lies flat on the operating table, and the body moves with the operating table to the detector. X-rays are used to scan a certain part of the body at a certain thickness. The operating table is usually made of metal, which is a high-density material. Its density is greater than that of human bones, so it will be highlighted under X-rays. Therefore, the table board is often visible in CT images.

[0003] In CT images, because the brightness of the bed board and human bones is extremely similar, the bed board is easily reconstructed when performing three-dimensional reconstruction of the bones, which can obscure certain areas of the bones and make it difficult to observe diseased bones and plan surgery.

[0004] Currently, the main methods for removing bed slabs from CT images are manual or semi-automatic operations by doctors. Specifically, the doctor selects the bed slab through an interactive interface, and then deletes it using region growing technology. However, this method has problems: doctors need to visually identify the bed slab and accurately select it with a mouse. Since some bed slabs are thin, the probability of selection is reduced, and the workload is heavy and can easily lead to fatigue. Therefore, automatic bed slab removal is of great significance for medical disease diagnosis.

[0005] In the prior art, the invention entitled "Method, System, Readable Storage Medium and Device for Removing Bed Board Based on CT Images" (application number 201911225985.7) mentions a method for removing the bed board, which relies on hardware to acquire CT images and requires acquiring empty bed board parameters to match the CT empty bed scan image with the CT scan image to be processed, thereby obtaining the bed-removed image of the CT scan image to be processed. This method is not suitable for common 3D imaging workstations. In the invention entitled "Method and System for Removing Bed Board from CT Images" (application number 201410438472.5), it is necessary to filter out edge curves with low similarity in the image and identify the edge curves of the bed board. This method relies too much on the physical characteristics of the bed and is prone to errors when filtering out interference.

[0006] To address the shortcomings of existing methods for removing bed slabs from CT images, this patented method employs a fully automated removal approach. It does not rely on hardware conditions or interference from edge curves similar to the bed slab, enabling rapid and accurate one-click removal. Summary of the Invention

[0007] To address the aforementioned problems and proposed solutions, this invention presents a method and system for automatically removing bed boards from CT images. By employing image segmentation technology, the bed boards in CT images can be quickly and automatically removed. This method is suitable for 3D model reconstruction in imaging workstations and can also prevent the bed boards from obscuring certain areas of the model, effectively meeting the needs of doctors and greatly improving work efficiency.

[0008] To achieve the above objectives, this invention discloses a method for automatically removing bed slabs from CT images, comprising the following steps:

[0009] Obtain the three-dimensional dimensions (X, Y, Z) of the CT image;

[0010] Construct a 2D image of size X and Y;

[0011] Traverse the pixel values ​​along the z-axis of the CT image, and assign the maximum pixel value at (x, y) to the (x, y) position in the 2D image.

[0012] The 2D image is thresholded based on the CT value to extract the bed board and human tissue regions. The pixel values ​​of the extracted regions and the remaining regions are set differently to obtain a binary image.

[0013] The binary image is subjected to seed point region growing processing to extract the human tissue region and obtain a mask_2D image;

[0014] Along the z-axis direction of the CT image, Z mask_2D images are stitched together to obtain a mask_3D image;

[0015] The CT image is segmented using the mask_3D image to obtain the bedRejected_3D image after removing the bed board.

[0016] As a further improvement of the present invention, the acquired CT image is first subjected to filtering and smoothing processing.

[0017] As a further improvement of the present invention, the initial pixel value of the 2D image is set to 0.

[0018] As a further improvement of the present invention, when performing threshold segmentation on the 2D image based on the CT value, a CT value lower than the CT value of human tissue and greater than the CT value of air is selected as the threshold.

[0019] As a further improvement of the present invention, when performing threshold segmentation on the 2D image based on the CT value, the pixel value of the extracted bed board and human tissue area is set to 1, and the pixel value of the remaining area is 0.

[0020] As a further improvement of the present invention, when performing seed point region generation processing on the binary image, the region growth threshold is set to 1.

[0021] As a further improvement of the present invention, when performing seed point region generation processing on the binary image, the seed points are set on the human tissue region.

[0022] As a further improvement of the present invention, when performing seed point region growing processing on the binary image, the pixel value of the extracted human tissue region is set to 1, and the pixel value of the remaining region is set to 0.

[0023] As a further improvement of the present invention, image template segmentation processing is performed on the CT image based on the mask_3D image, including:

[0024] The mask_3D image is used as an image segmentation template;

[0025] Set the foreground value to the pixel value of the CT image and the background value to the CT value of the air;

[0026] Obtain the bedRejected_3D image after removing the bed board.

[0027] The present invention also provides a system for automatic removal of bed boards in CT images, comprising: a 2D image construction module, a threshold segmentation extraction module, a region growing extraction module, an image segmentation template acquisition module, and an image template segmentation processing module;

[0028] The 2D image construction module is used for:

[0029] Obtain the three-dimensional dimensions (X, Y, Z) of the CT image;

[0030] Construct a 2D image of size X and Y;

[0031] Traverse the pixel values ​​along the z-axis of the CT image, and assign the maximum pixel value at (x, y) to the (x, y) position in the 2D image.

[0032] The threshold segmentation and extraction module is used for:

[0033] The 2D image is thresholded based on the CT value to extract the bed board and human tissue regions. The pixel values ​​of the extracted regions and the remaining regions are set differently to obtain a binary image.

[0034] The region growth extraction module is used for:

[0035] The binary image is subjected to seed point region growing processing to extract the human tissue region and obtain a mask_2D image;

[0036] The image segmentation template acquisition module is used for:

[0037] Along the z-axis direction of the CT image, Z mask_2D images are stitched together to obtain a mask_3D image;

[0038] The image template segmentation processing module is used for:

[0039] The CT image is segmented using the mask_3D image to obtain the bedRejected_3D image after removing the bed board.

[0040] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0041] This invention utilizes image segmentation technology to quickly and automatically remove bed boards from CT images, achieving fully automated bed board removal. Compared to existing technologies where doctors manually or semi-automatically remove bed boards, this method saves manpower and time, improves work efficiency, and meets doctors' image processing needs.

[0042] This invention cleverly uses two image extraction methods. First, threshold segmentation is used to extract the human tissue region and the bed board, with a clear interval between them. Then, the seed point region growth method is used to extract the human tissue region, achieving perfect removal of the bed board. Attached Figure Description

[0043] Figure 1 This is a flowchart of a method for automatically removing bed boards from CT images according to an embodiment of the present invention;

[0044] Figure 2 This is a schematic diagram of a system for automatic removal of bed boards from CT images, as disclosed in one embodiment of the present invention.

[0045] Figure 3 This is a schematic cross-sectional view of a CT scan image from hip to ankle of a patient with knee joint disease, as disclosed in an embodiment of the present invention.

[0046] Figure 4 for Figure 3 A schematic diagram of the binary image obtained after thresholding of the image;

[0047] Figure 5 for Figure 4 A schematic diagram illustrating the selection of seed points in a binary image;

[0048] Figure 6 Therefore Figure 5 The mask_2D image obtained after region growing from seed points;

[0049] Figure 7A mask 3D image is obtained by stitching together Z mask 2D images;

[0050] Figure 8 This is a schematic diagram of the bedRejected_3D image obtained by using the mask_3D image as an image segmentation template to perform image template segmentation on the CT image. Detailed Implementation

[0051] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0052] The present invention will now be described in further detail with reference to the accompanying drawings:

[0053] like Figure 1 As shown, the present invention discloses a method for automatic removal of bed slabs from CT images, comprising:

[0054] S1. Obtain the patient's CT images through scanning, and obtain the three-dimensional dimensions X, Y, and Z of the CT images;

[0055] in,

[0056] The acquired CT images need to be filtered and smoothed first, such as... Figure 2 The image shown is a cross-sectional image of a full-length CT scan from hip to ankle of a patient with knee joint disease, after being filtered and smoothed.

[0057] S2. Construct a 2D image of size X and Y;

[0058] in,

[0059] Set the initial pixel value of the 2D image to 0.

[0060] S3. Traverse the pixel values ​​along the z-axis of the CT image and assign the maximum pixel value at (x, y) in the 2D image to the value at (x, y).

[0061] S4. Threshold segment the 2D image based on the CT value to extract the bed board and human tissue regions. Differentiate the pixel values ​​of the extracted and remaining regions to obtain a binary image, such as... Figure 4 As shown;

[0062] in,

[0063] Since there is air and mattress between the bed board and the human body, a CT value lower than the CT value of human tissue and higher than the CT value of air is selected as the threshold, preferably -850.

[0064] Set the pixel value of the extracted bed board and human tissue area to 1, and the pixel value of the remaining area (i.e. the air and mattress area between the bed board and human tissue area) to 0.

[0065] S5. Perform seed point region growing processing on the binary image to extract the human tissue region and obtain the mask_2D image, such as... Figure 6 As shown;

[0066] in,

[0067] The seed point is set on the human tissue area, preferably at the center of the human tissue. For example... Figure 5 As shown;

[0068] The region growth threshold is set to 1 because the pixel value of the human tissue region is 1. This allows us to merge adjacent pixels with similar pixel values ​​to each seed point into the same region, starting from the seed point. Growth stops when the pixel value of an adjacent pixel is not 1. Since the pixel value of the region between the human tissue region and the bed board is 0, growth stops when the human tissue region reaches its boundary.

[0069] Set the pixel values ​​of the extracted human tissue regions to 1, and the pixel values ​​of the remaining regions to 0, to obtain a mask_2D image containing only human tissue regions.

[0070] S6. Along the z-axis of the CT image, stitch together the Z mask_2D images to obtain the mask_3D image, as shown below. Figure 7 As shown;

[0071] The mask_3D image is a columnar structure formed by multiplying the maximum area of ​​the human tissue region along the z-axis.

[0072] S7. Perform image template segmentation on the CT image based on the mask_3D image to obtain the bedRejected_3D image after removing the bed board, as shown below. Figure 8 As shown.

[0073] The process of image template segmentation of CT images based on the mask_3D image includes:

[0074] Use the mask_3D image as an image segmentation template;

[0075] Set the foreground value to the pixel value of the CT image and the background value to the CT value of air, which is -1024.

[0076] Obtain the bedRejected_3D image after removing the bed board.

[0077] like Figure 2 As shown, the present invention also provides a system for automatic removal of bed boards in CT images, comprising: a 2D image construction module, a threshold segmentation extraction module, a region growing extraction module, an image segmentation template acquisition module, and an image template segmentation processing module;

[0078] 2D image building module, used for:

[0079] Obtain the three-dimensional dimensions (X, Y, Z) of the CT image;

[0080] Construct a 2D image of size X and Y;

[0081] Traverse the pixel values ​​along the z-axis of the CT image and assign the maximum pixel value at (x, y) in the 2D image to the value at (x, y).

[0082] The threshold segmentation and extraction module is used for:

[0083] Thresholding is performed on the 2D image based on the CT value to extract the bed board and human tissue regions. The pixel values ​​of the extracted regions and the remaining regions are set differently to obtain a binary image.

[0084] The region growth extraction module is used for:

[0085] Seed point region growing processing is performed on the binary image to extract the human tissue region and obtain the mask_2D image;

[0086] The image segmentation template acquisition module is used for:

[0087] Along the z-axis of the CT image, Z mask_2D images are stitched together to obtain a mask_3D image;

[0088] The image template segmentation processing module is used for:

[0089] The CT image is segmented using the mask_3D image to obtain the bedRejected_3D image after removing the bed board.

[0090] Advantages of this invention:

[0091] (1) By using image segmentation technology, the bed board in the CT image can be removed automatically and quickly, realizing the fully automatic removal of the bed board. Compared with the existing technology center's manual and semi-automatic methods for removing the bed board, it saves manpower and time, improves work efficiency, and meets the image processing needs of doctors.

[0092] (2) The present invention cleverly uses two image extraction methods. First, the human tissue area and the bed board are extracted by threshold segmentation, and there is a clear gap between them. Then, the human tissue area is extracted by seed point region growth method, so as to achieve perfect removal of the bed board.

[0093] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for automatically removing bed slabs from CT images, characterized in that, Including the following steps: Obtain the three-dimensional dimensions (X, Y, Z) of the CT image; Construct a 2D image of size X and Y; Traverse the pixel values ​​along the z-axis of the CT image, and assign the maximum pixel value at (x, y) to the (x, y) position in the 2D image. The 2D image is thresholded based on the CT value to extract the bed board and human tissue regions. The pixel values ​​of the extracted regions and the remaining regions are set differently to obtain a binary image. The binary image is subjected to seed point region growing processing to extract the human tissue region and obtain a mask_2D image; Along the z-axis direction of the CT image, Z mask_2D images are stitched together to obtain a mask_3D image; The CT image is segmented using the mask_3D image to obtain the bedRejected_3D image after removing the bed board.

2. The method according to claim 1, characterized in that: The acquired CT images are first subjected to filtering and smoothing processing.

3. The method according to claim 1, characterized in that: Set the initial pixel value of the 2D image to 0.

4. The method according to claim 1, characterized in that: When performing threshold segmentation on the 2D image based on the CT value, a CT value lower than the CT value of human tissue but greater than the CT value of air is selected as the threshold.

5. The method according to claim 1, characterized in that: When performing threshold segmentation on the 2D image based on CT values, the pixel values ​​of the extracted bed board and human tissue regions are set to 1, and the pixel values ​​of the remaining regions are set to 0.

6. The method according to claim 5, characterized in that: When performing seed point region generation processing on the binary image, the region growth threshold is set to 1.

7. The method according to claim 1, characterized in that: When performing seed point region generation processing on the binary image, the seed points are set on the human tissue region.

8. The method according to claim 1, characterized in that: When performing seed point region growing processing on the binary image, the pixel values ​​of the extracted human tissue regions are set to 1, and the pixel values ​​of the remaining regions are set to 0.

9. The method according to claim 1, characterized in that: Image template segmentation processing of the CT image based on the mask_3D image includes: The mask_3D image is used as an image segmentation template; Set the foreground value to the pixel value of the CT image and the background value to the CT value of the air; Obtain the bedRejected_3D image after removing the bed board.

10. A system based on the method according to any one of claims 1 to 9, characterized in that, include: The module includes a 2D image construction module, a threshold segmentation and extraction module, a region growing and extraction module, an image segmentation template acquisition module, and an image template segmentation processing module. The 2D image construction module is used for: Obtain the three-dimensional dimensions (X, Y, Z) of the CT image; Construct a 2D image of size X and Y; Traverse the pixel values ​​along the z-axis of the CT image, and assign the maximum pixel value at (x, y) to the (x, y) position in the 2D image. The threshold segmentation and extraction module is used for: The 2D image is thresholded based on the CT value to extract the bed board and human tissue regions. The pixel values ​​of the extracted regions and the remaining regions are set differently to obtain a binary image. The region growth extraction module is used for: The binary image is subjected to seed point region growing processing to extract the human tissue region and obtain a mask_2D image; The image segmentation template acquisition module is used for: Along the z-axis direction of the CT image, Z mask_2D images are stitched together to obtain a mask_3D image; The image template segmentation processing module is used for: The CT image is segmented using the mask_3D image to obtain the bedRejected_3D image after removing the bed board.

Images