An automatic segmentation method for complex bone models integrating skeleton lines

Abstract

The present invention discloses an automatic segmentation method for a complex bone model fused with skeleton lines, comprising the following steps: Step 1: Determine the boundary feature points of the bone template segmentation. Construct a geometric average model as a template, interactively select the segmentation boundary based on the medical definition of the bone, and determine the boundary feature points on the segmentation boundary according to the surface geometric properties; Step 2: Select the target bone band segmentation area. Extract the skeleton lines of the bone template and the target bone respectively, and determine the band segmentation area of ​​the target bone according to the proportional division of the skeleton lines in the template; Step 3: Accurate segmentation of the target bone. According to the geometric properties of the template boundary feature points, the boundary feature points are determined in the band segmentation area of ​​the target bone, and the segmentation curve is obtained from the boundary feature points to achieve the segmentation of the target bone. The present invention can quickly segment complex bones into different areas that are convenient for research, save time and labor costs, and is of great significance for the judgment of the type of bone damage.

Description

Technical Field

[0001] The invention relates to a digital orthopedic technology, and in particular to an automatic segmentation method for a complex bone model fused with skeleton lines. The invention belongs to the field of computer-aided design. Background Art

[0002] The rise of digital orthopedics in clinical applications has gradually transformed the way doctors diagnose and treat clinical orthopedic patients, and has a significant effect on improving the scientific nature of orthopedic surgical treatment. Many bones in the human body are complex in shape. Take the spine as an example. The spine is the bone that supports the human body and is closely related to human life. At present, due to long-term use of electronic products and incorrect sitting posture, spinal diseases have become a very common disease in the Chinese population, and show a trend of spreading to young people. Segmenting bones can divide complex bones into relatively simple modules, which is convenient for doctors to study bones. In addition, bone segmentation helps to judge the type of damaged bones in order to achieve the purpose of accurate treatment of patients with bone damage.

[0003] However, the shape of the vertebra is irregular, the surface is a free-form surface and has multiple protrusions, so it is difficult to study it directly. In order to reduce the complexity of the skeleton, the skeleton is divided into multiple areas of medical significance by segmentation, and each area is studied separately. The current segmentation method is mainly manual segmentation by doctors with the assistance of medical images, which depends to a large extent on the experience and professional level of the doctors. Different doctors have different standards for segmentation and have a certain degree of subjectivity. In addition, due to differences in age, gender, etc., there are differences in individual bone morphology, and manual segmentation is time-consuming and labor-intensive. Therefore, we hope to be able to use computers to quickly and effectively realize automatic segmentation of bones. The present invention proposes to merge the skeleton line with the template, use the skeleton line on the template to determine the segmentation band area of ​​the target bone, and then determine the segmentation boundary feature points through geometric properties to achieve the segmentation of the target bone. The present invention aims to reduce the subjectivity of manual operation and assist doctors in studying complex bones, and is of great significance to the treatment of patients with bone damage. Summary of the invention

[0004] The present invention proposes an automatic segmentation method for complex bone models fused with skeleton lines, which can solve the problem of difficult research on complex bones, reduce the burden on doctors, and help determine the type of bone damage. In order to achieve the above goals, the present invention adopts the following technical solutions:

[0005] A complex skeleton model automatic segmentation method integrating skeleton lines comprises the following steps:

[0006] Step 1: Determine the boundary feature points of the bone template segmentation;

[0007] A geometric average model is constructed as a template, segmentation boundaries are interactively selected based on the medical definition of bones, and boundary feature points are determined on the segmentation boundaries according to surface geometric properties;

[0008] Step 1a: Use a large number of bones to build a geometric average model, reconstruct the geometric average model data into a three-dimensional point cloud model as a template, interactively select segmentation boundaries according to the bone medical definition, and divide the template into multiple regions;

[0009] In step 1a, the interactive selection of segmentation boundaries means that after a line segment is drawn using a mouse, the computer automatically selects points on the point cloud model whose distances to the line segment are less than a set threshold and connects them into a curve.

[0010] Step 1b: On the segmentation boundary obtained in step 1a, according to the geometric properties of the model surface, select the point with the most obvious geometric properties as the segmentation boundary feature point, where the most obvious geometric property is defined as the curvature change greater than the set threshold;

[0011] Step 2: Select the target bone strip segmentation area;

[0012] Extract the skeleton lines of the skeleton template and the target skeleton respectively, and determine the strip segmentation area of ​​the target skeleton according to the proportion of the skeleton lines in the template;

[0013] Step 2a: Mark a significant feature point in each area of ​​the template, and extract the skeleton line based on the significant feature point. The skeleton line intersects with the segmentation surface formed by the segmentation boundary at a point, which divides the skeleton line into two continuous segments, and obtains the ratio of the two skeleton lines;

[0014] In steps 2a and 2b, the skeleton line is extracted based on the significant feature points. For two adjacent regions, the skeleton line is extracted using the maximum inscribed sphere method starting from the significant feature points of one region and ending at the significant feature points of the other region, while for non-adjacent regions, the skeleton line is not extracted.

[0015] Step 2b: Similar to step 2a, the significant feature points are marked on the target skeleton, the skeleton line is extracted, the segmentation point of the target skeleton line is obtained according to the ratio of the template skeleton line, and the strip segmentation area is divided with the segmentation point as the center;

[0016] In step 2b, the division of the strip segmentation area with the skeleton line segmentation point as the center means that in the area where the skeleton line passes, starting from the segmentation point, there is an angle of The direction of the beam, where For a fixed angle, a group of rays are drawn along this direction, and the rays intersect with the model surface, forming a strip-shaped segmentation area at the intersection position.

[0017] Step 3: Accurate segmentation of target bones;

[0018] According to the geometric properties of the template boundary feature points, the boundary feature points are determined in the strip segmentation area of ​​the target bone, and the segmentation curve is obtained from the boundary feature points to achieve the segmentation of the target bone.

[0019] Step 3a: According to the geometric properties of the boundary feature points on the template, points with the same geometric properties are selected as boundary feature points of the target bone in the strip segmentation area of ​​the target bone in step 2b;

[0020] Step 3b: For the boundary feature points selected in step 3a, an interpolation method is used to obtain a closed curve on the surface of the target bone, and the target bone is segmented by the curve.

[0021] The benefit of the present invention lies in that the present invention is applied in the field of medical orthopedics and can help doctors study complex bones, thereby accurately determining the type of bone damage, which is of great significance for the treatment of damaged bones. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 This is the process of the automatic segmentation method of the complex skeleton model fused with skeleton lines in the present invention;

[0023] Figure 2 It is the partial segmentation result of the skeleton template;

[0024] Figure 3 It is the boundary feature point of the skeleton template segmentation;

[0025] Figure 4 It is the process of forming the target bone band segmentation area;

[0026] Figure 5 It is the process of forming the segmentation curve of the target bone. DETAILED DESCRIPTION

[0027] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0028] like Figure 1 As shown, a complex skeleton model automatic segmentation method integrating skeleton lines includes the following steps:

[0029] Step 1: Determine the boundary feature points of the bone template segmentation;

[0030] A geometric average model is constructed as a template, segmentation boundaries are interactively selected based on the medical definition of bones, and boundary feature points are determined on the segmentation boundaries according to surface geometric properties;

[0031] Step 1a: Use a large number of bones to build a geometric average model, reconstruct the geometric average model data into a 3D point cloud model as a template, interactively select segmentation boundaries according to the bone medical definition, and divide the template into multiple regions, such as Figure 2 As shown;

[0032] The aforementioned interactive selection of segmentation boundaries means that after a line segment is drawn using a mouse, the computer automatically selects points on the point cloud model whose distance to the line segment is less than a set threshold and connects them into a curve;

[0033] Step 1b: On the segmentation boundary obtained in step 1a, according to the geometric properties of the model surface, select the point with the most obvious geometric properties as the segmentation boundary feature point. The most obvious geometric property is defined as the curvature change greater than the set threshold, such as Figure 3 As shown;

[0034] Step 2: Select the target bone strip segmentation area;

[0035] Extract the skeleton lines of the skeleton template and the target skeleton respectively, and determine the strip segmentation area of ​​the target skeleton according to the proportion of the skeleton lines in the template;

[0036] Step 2a: Mark a significant feature point in each area of ​​the template, and extract the skeleton line based on the significant feature point. The skeleton line intersects with the segmentation surface formed by the segmentation boundary at a point, which divides the skeleton line into two continuous segments, and obtains the ratio of the two skeleton lines, such as Figure 4 (a)

[0037] Step 2b: Similar to step 2a, the salient feature points are marked on the target skeleton, the skeleton line is extracted, the segmentation point of the target skeleton line is obtained according to the ratio of the template skeleton line, and the strip segmentation area is divided with the segmentation point as the center, such as Figure 4 As shown, Figure 4 (d) is the strip segmentation area of ​​the target bone finally obtained according to the method;

[0038] The aforementioned extraction of skeleton lines based on significant feature points means that for two adjacent regions, the skeleton line is extracted using the maximum inscribed sphere method starting from the significant feature points of one region and ending at the significant feature points of the other region, while for non-adjacent regions, no skeleton line extraction is performed;

[0039] The aforementioned division of the strip segmentation area with the skeleton line segmentation point as the center means that in the area where the skeleton line passes, starting from the segmentation point, there is an angle of The direction of the beam, where is a fixed angle, and a set of rays are drawn along this direction. The rays intersect with the model surface, such as Figure 4As shown in (b), the intersecting positions form a strip-shaped segmentation area, such as Figure 4 (c)

[0040] Step 3: Accurate segmentation of target bones;

[0041] According to the geometric properties of the template boundary feature points, the boundary feature points are determined in the strip segmentation area of ​​the target bone, and the segmentation curve is obtained from the boundary feature points to achieve the segmentation of the target bone.

[0042] Step 3a: Based on the geometric properties of the boundary feature points on the template, the target bone strip segmentation area (i.e. Figure 5 (a)), select points with the same geometric properties as the boundary feature points of the target bone, such as Figure 5 (b) shows the boundary feature points selected in the strip segmentation area;

[0043] Step 3b: For the boundary feature points selected in step 3a, an interpolation method is used to obtain a closed curve on the target bone surface, such as Figure 5 (c) shows the segmentation curve formed in the strip segmentation area. Figure 5 (d) is the final segmentation curve, which is used to segment the target skeleton;

[0044] Anything not further described in the present invention is prior art.

[0045] The present invention provides a method for automatic segmentation of complex bone models fused with skeleton lines. First, a geometric average model is constructed as a template, and the boundary feature points are determined after selecting the segmentation boundary based on the medical definition of the bone; then, the skeleton lines of the bone template and the target bone are extracted respectively, and the band segmentation area of ​​the target bone is determined according to the proportion of the skeleton lines in the template; finally, the boundary feature points are determined in the band segmentation area, and the segmentation curve is formed by the boundary feature points. The present invention can not only quickly and accurately segment complex bones, but also has important significance for judging the type of bone damage.

[0046] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A complex skeleton model automatic segmentation method integrating skeleton lines, characterized in that: The steps include: Step 1: Determine the boundary feature points of the bone template segmentation; A geometric average model is constructed as a template, segmentation boundaries are interactively selected based on the medical definition of bones, and boundary feature points are determined on the segmentation boundaries according to surface geometric properties; Step 1a: Use a number of bones to construct a geometric average model, reconstruct the geometric average model data into a three-dimensional point cloud model as a template, interactively select segmentation boundaries according to the bone medical definition, and divide the template into multiple regions; Step 1b: On the segmentation boundary obtained in step 1a, according to the geometric properties of the model surface, select the point with the most obvious geometric properties as the segmentation boundary feature point, wherein the most obvious geometric property is defined as the curvature change greater than a set threshold; Step 2: Select the target bone strip segmentation area; Extract the skeleton lines of the skeleton template and the target skeleton respectively, and determine the strip segmentation area of ​​the target skeleton according to the proportion of the skeleton lines in the template; Step 2a: Mark a significant feature point in each area of ​​the template, and extract a skeleton line based on the significant feature point; the skeleton line intersects with the segmentation surface formed by the segmentation boundary at a point, and the point divides the skeleton line into two continuous segments, and obtains the ratio of the two skeleton lines; Step 2b: Similar to step 2a, the significant feature points are marked on the target skeleton, the skeleton line is extracted, the segmentation point of the target skeleton line is obtained according to the ratio of the template skeleton line, and the strip segmentation area is divided with the segmentation point as the center; Step 3: Accurate segmentation of target bones; According to the geometric properties of the template boundary feature points, the boundary feature points are determined in the strip segmentation area of ​​the target bone, and the segmentation curve is obtained from the boundary feature points to achieve the segmentation of the target bone; Step 3a: According to the geometric properties of the boundary feature points on the template, points with the same geometric properties are selected as boundary feature points of the target bone in the strip segmentation area of ​​the target bone in step 2b; Step 3b: For the boundary feature points selected in step 3a, an interpolation method is used to obtain a closed curve on the surface of the target bone, and the target bone is segmented by the curve.

2. The complex skeleton model automatic segmentation method of claim 1, characterized in that: In step 1a, the interactive selection of segmentation boundaries means that after a line segment is drawn using a mouse, the computer automatically selects points on the point cloud model whose distances to the line segment are less than a set threshold and connects them into a curve.

3. The complex skeleton model automatic segmentation method of claim 1, characterized in that: In steps 2a and 2b, the skeleton line is extracted based on the significant feature points. For two adjacent regions, the skeleton line is extracted using the maximum inscribed sphere method starting from the significant feature points of one region and ending at the significant feature points of the other region, while for non-adjacent regions, the skeleton line is not extracted.

4. The complex skeleton model automatic segmentation method of claim 1, characterized in that: In step 2b, the strip segmentation area is divided with the skeleton line segmentation point as the center, which means that in the area where the skeleton line passes, there is a direction beam starting from the segmentation point and having an angle θ with the plane where the normal vector of the segmentation point on the skeleton line is located, wherein θ is a fixed angle, and a group of rays are made along the direction beam, the rays intersect with the model surface, and a strip segmentation area is formed by the intersection position.

Images