System and method for acetabulum and femur segmentation and femoral head identification

Abstract

The present invention relates to the technical field of medical image processing, and provides a system for acetabulum and femur segmentation and femoral head identification, the system comprising a memory, an image acquisition apparatus and a processor. The processor comprises: an image reading module, configured to read an acetabulum-femur image; an image segmentation module, configured to use a region growing algorithm to perform segmentation of the acetabulum-femur image on the basis of a pixel value range of pixels, wherein a region formed by pixels within the pixel value range is an acetabular position segmentation map or a femoral position segmentation map; and a femoral head identification module, configured to use a femur identification algorithm to identify the femoral head on the basis of the femoral position segmentation map, and calculate the sphere center position and the maximum radius of the femoral head. The system involves a clear and simple execution process, allows users to automatically adjust input parameters of the algorithms on the basis of data, so as to be suitable for lesion images of different types, and, while achieving the same effects, reduces the development costs, improves the development efficiency, and ensures rapid verification for preoperative planning.

Description

A system and method for segmenting the acetabular femur and identifying the femoral head. Technical Field

[0001] This invention relates to the field of medical image processing technology, and more particularly to a system and method for segmenting the acetabulum and identifying the femoral head. Background Technology

[0002] In image-guided hip arthroscopic minimally invasive surgery, it is usually necessary to preprocess the preoperatively acquired 3D CT images, segment the acetabulum and femur, locate the center of the femoral head, and calculate its maximum radius. This is then used for subsequent preoperative planning of the grinding area for intraoperative navigation.

[0003] In existing technologies, the mainstream approach for identifying the femoral head of the hip joint is to segment the acetabulum and femur, then select a femoral head region, obtain the set of most points in the femoral head region, and then use a spherical fitting algorithm to identify the femoral head and calculate the position and radius of the sphere center.

[0004] However, existing technologies typically encounter two problems. One problem is that when encountering actual lesion images with uneven bone density distribution and narrow areas of the acetabulum and femoral head, traditional image processing algorithms cannot successfully segment the acetabulum and femur. On the other hand, deep learning segmentation algorithms require a large number of training samples and manual annotation, resulting in low time consumption and efficiency. The other problem is that when performing spherical fitting of the femur after segmentation, since the femoral head data is not actually distributed on a sphere, there is a significant difference between the fitting result and the actual location of the femur, and the fitting algorithm is also time-consuming. Summary of the Invention

[0005] To address the aforementioned problems, the present invention aims to provide a system and method for segmenting the acetabulum and identifying the femoral head. The system has a clear and simple execution process, obvious effects, and the algorithm's input parameters can be automatically adjusted by the user according to the data. It is applicable to different types of lesion images, and while achieving the same effect, it saves development costs, improves development efficiency, and provides a guarantee for rapid verification of subsequent operations such as preoperative planning.

[0006] The above-mentioned objective of this invention is achieved through the following technical solutions:

[0007] A system for segmenting the acetabular femur and identifying the femoral head includes:

[0008] The memory stores the acquired acetabular-femoral images that require acetabular-femoral segmentation and femoral head identification, as well as the segmented acetabular position map, femoral position map, and the calculated center position and maximum radius of the femoral head.

[0009] An image acquisition device for acquiring images of the acetabulum and femur, including acetabular and femoral tissue, before image-guided hip arthroscopic minimally invasive surgery;

[0010] The processor is used in the image-guided arthroscopic minimally invasive surgery to segment the acetabular femur from preoperative images of the acetabular femur showing diseased bone tissue, and to identify the femoral head for navigation during the subsequent image-guided arthroscopic minimally invasive surgery. The processor segments the acetabular femur and identifies the femoral head through the following modules:

[0011] The image reading module is used to read the acetabular-femoral image that needs to be segmented and the femoral head identified, wherein the acetabular-femoral image is an image including the acetabulum and femoral tissue acquired by the image acquisition device before the image-guided hip arthroscopy minimally invasive surgery.

[0012] The image segmentation module is used to segment the acetabular femur image based on the range of pixel values ​​using a region growing algorithm. The region enclosed by the pixels within the range of pixel values ​​is the acetabular position segmentation map or the femur position segmentation map.

[0013] The femoral head recognition module is used to identify the femoral head based on the femoral position segmentation map and the femoral head recognition algorithm, and to calculate the center position and maximum radius of the femoral head.

[0014] Furthermore, the processor also includes a gradient image calculation module;

[0015] The gradient image calculation module is used to calculate the gradient of the acetabular femur image before segmenting the acetabular femur using the image segmentation module. When the gap between the femur and acetabulum in the read acetabular femur image is narrow, or the pixel value difference between the non-target region and the target region is not significant, the gradient image of the acetabular femur image is obtained, and the gradient image is superimposed on the original acetabular femur image. The target region is the region where the acetabulum or the femur is located, and the non-target region is the region other than the region where the acetabulum and the femur are located.

[0016] Further, in the gradient image calculation module, the gradient of the acetabular femoral image is calculated to obtain the gradient image of the acetabular femoral image, and the gradient image is superimposed on the original acetabular femoral image, specifically as follows:

[0017] The gradient image is obtained using a Gaussian function for image filtering and edge detection;

[0018] The Gaussian function parameter σ is selected based on data including the image noise level, image resolution, image contrast and brightness, and the size of the target region of the acetabular femur image. The higher the image noise level, the larger the Gaussian function parameter σ; the lower the image resolution, the larger the Gaussian function parameter σ; the lower the image contrast and brightness, the larger the Gaussian function parameter σ; and the smaller the target region, the larger the Gaussian function parameter σ.

[0019] Substitute the Gaussian function parameter σ into the Gaussian function to calculate the first-order partial derivative of the Gaussian function, thereby obtaining the first-order gradient function of the acetabular femur image;

[0020] Discretize the first-order gradient function to obtain a first-order gradient operator. Convolve the first-order gradient operator with the acetabular femur image and calculate the gradient magnitude of each pixel to obtain the gradient image representing the gradient magnitude of the image.

[0021] Furthermore, in the image segmentation module, the region growing algorithm is used to segment the acetabular-femoral image based on the range of pixel values. The region enclosed by the pixels within the range of pixel values ​​is the acetabular position segmentation map or the femoral position segmentation map, specifically:

[0022] Click n+1 pixels on the acetabular femur image as seed points. The first n seed points are on the target area, and the (n+1)th seed point is on the non-target area.

[0023] The upper and lower bound thresholds for segmenting the acetabular femur image are calculated using the pixel values ​​of the n+1 seed points clicked.

[0024] Starting from the seed point, a breadth-first search is performed on the acetabular femur image using the upper and lower bound thresholds as boundaries to obtain the target segmentation result. The breadth-first search is an algorithm that traverses the image storage structure, sequentially traversing each vertex in the acetabular femur image from six directions: top, bottom, left, right, front, and back, to obtain all pixels that satisfy the upper and lower bound thresholds as the target segmentation result. The image corresponding to the target segmentation result is either the acetabular position segmentation map or the femur position segmentation map.

[0025] Further, in the image segmentation module, the upper and lower bound thresholds for segmenting the acetabular femur image are calculated based on the pixel values ​​of the clicked n+1 seed points, specifically as follows:

[0026] Calculate the average pixel value of the first n seed points;

[0027] The average pixel value of the first n seed points is added to a first preset value to serve as the upper bound threshold.

[0028] The average pixel value of the first n seed points is subtracted from the second preset value to obtain the lower bound preset value. The lower bound preset value is compared with the pixel value of the (n+1)th seed point, and the larger value between the lower bound preset value and the pixel value of the (n+1)th seed point is taken as the lower bound threshold.

[0029] Furthermore, the image segmentation module further includes: smoothing the edges of the target segmentation result, specifically:

[0030] The edge regions of the target segmentation result are obtained using edge detection algorithms including Sobel and Canny.

[0031] The acquired edge region is dilated to expand the edge region by a preset number of pixels, ensuring that subsequent smoothing operations do not lose the edge information of the target segmentation result.

[0032] The expanded edge region is smoothed by applying smoothing filters, including Gaussian filtering and bilateral filtering, to reduce edge noise and discontinuities, thus making the edges smooth.

[0033] An erosion operation is performed on the smoothed edge region to restore the edge region to its original size, retain the main information of the edge, and remove unnecessary pixels introduced during the smoothing process.

[0034] The smoothed edge region is merged with the original target segmentation result to obtain the final smoothed target segmentation result.

[0035] Furthermore, in the femoral head recognition module, the position of the ball of the foot is calculated as follows:

[0036] In the image segmentation module, the pixel values ​​of the pixels in the segmented femoral position segmentation image are marked with 1 and 0, where 1 represents the femoral bone and 0 represents the non-target region.

[0037] Take slice data from the positive z-axis direction towards the origin, sum the pixel values ​​in the slice data, and when the sum of the pixel values ​​of the first slice data is not 0, the average of the x-axis coordinate values ​​of the non-zero pixels in the slice data is the x-coordinate of the center of the sphere, denoted as x, and the average of the y-axis coordinate values ​​of the non-zero pixels in the slice data is the y-coordinate of the center of the sphere, denoted as y. At the same time, the z-axis coordinate of the slice data at this time is denoted as z1.

[0038] Take slice data from the positive x-axis direction toward the origin, sum the pixel values ​​in the slice data, and when the sum of the pixel values ​​of the first slice data is not 0, the average value of the z-axis coordinates of the non-zero pixels in the slice data is the z-coordinate of the center of the sphere, denoted as z, and the x-axis coordinate of the slice data at this time is denoted as x1.

[0039] Further, in the image segmentation module, the maximum radius is calculated as follows:

[0040] Find the maximum value among z1-z and x1-x, where the maximum value is the maximum radius.

[0041] A method for segmenting the acetabular femur and identifying the femoral head using the system described above includes:

[0042] S1: Read the acetabular-femoral image that needs to be segmented and the femoral head identified, wherein the acetabular-femoral image is an image of the acetabulum and femoral tissue acquired by medical imaging equipment before the image-guided hip arthroscopy minimally invasive surgery.

[0043] S2: Using a region growing algorithm, the acetabular femur image is segmented according to the range of pixel values. The region enclosed by the pixels within the range of pixel values ​​is used to obtain the acetabular position segmentation map or the femur position segmentation map.

[0044] S3: Based on the femoral position segmentation map, a femoral head recognition algorithm is used to identify the femoral head and calculate the center position and maximum radius of the femoral head.

[0045] A computer-readable storage medium storing computer code that, when executed, performs the method described above.

[0046] Compared with the prior art, the present invention has at least one of the following beneficial effects:

[0047] (1) A system for segmenting the acetabular femur and identifying the femoral head is provided, comprising: a memory storing acquired images of the acetabular femur requiring segmentation and femoral head identification, as well as segmented acetabular position maps, femoral position maps, calculated center position and maximum radius of the femoral head; an image acquisition device for acquiring images of the acetabular femur including acetabular and femoral tissues preoperatively for image-guided arthroscopic minimally invasive surgery; and a processor for segmenting the acetabular femur from the preoperatively acquired images of the acetabular femur showing diseased bone tissue and identifying the femoral head during the image-guided arthroscopic minimally invasive surgery for navigation in subsequent procedures, and by using... The following modules are used for segmenting the acetabulum and femur and identifying the femoral head: An image reading module is used to read the image of the acetabulum and femur that needs to be segmented and the femoral head identified. This image is acquired preoperatively via an image acquisition device, including the acetabulum and femoral tissue. An image segmentation module is used to segment the acetabulum and femur image using a region growing algorithm, based on the range of pixel values. The region enclosed by the pixels within the range of pixel values ​​is the acetabular or femoral position segmentation map. A femoral head identification module is used to identify the femoral head based on the femoral position segmentation map using a femoral head identification algorithm, and to calculate the center position and maximum radius of the femoral head. This technical solution is a simplified process for segmenting the acetabulum and femur and identifying the femoral head. It enables rapid segmentation of the acetabulum and femur and rapid location of the femoral position and size, with the advantages of short processing time and high effectiveness.

[0048] (2) When the gap between the femur and acetabulum is narrow, or the pixel value difference between the non-target region and the target region is not significant, the gradient of the acetabular-femoral image is calculated to obtain the gradient image of the acetabular-femoral image. The gradient image is then superimposed onto the original acetabular-femoral image. The target region is the area where the acetabulum or femur is located, and the non-target region is the area other than the acetabulum and femur. This technical solution, for actual lesion images with uneven bone density distribution and images where the acetabulum and femoral head regions are relatively narrow, can effectively segment the acetabulum and femur, as well as the target and non-target regions, by superimposing a gradient image onto the original image. Attached Figure Description

[0049] Figure 1 is a diagram of the overall algorithm executed by the processor in the system for segmenting the acetabulum and identifying the femoral head according to the present invention.

[0050] Figure 2 is a flowchart of the gradient calculation algorithm of the present invention;

[0051] Figure 3 is a schematic diagram of the gradient image of the present invention;

[0052] Figure 4 is a flowchart of the region growth algorithm of the present invention;

[0053] Figure 5 is a segmentation diagram of the femoral position in this invention;

[0054] Figure 6 is a segmentation diagram of the acetabulum position according to the present invention;

[0055] Figure 7 is a flowchart of the femoral bone recognition algorithm of the present invention, wherein (a) and (b) are flowcharts and schematic diagrams for calculating the x and y coordinates of the sphere center, and (c) and (d) are flowcharts and schematic diagrams for calculating the z coordinate of the sphere center;

[0056] Figure 8 shows the results of the femoral head bulb center and radius of the present invention;

[0057] Figure 9 is an overall flowchart of the method for segmenting the acetabulum and identifying the femoral head according to the present invention. Detailed Implementation

[0058] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0059] Those skilled in the art will understand that, unless specifically stated otherwise, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms. It should be further understood that the term “comprising” as used in this specification means the presence of the stated features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0060] First Embodiment

[0061] This embodiment provides a system for segmenting the acetabulum and femur and identifying the femoral head. After reading the acetabulum and femur image, users can customize whether to calculate the image gradient and whether to overlay the gradient image onto the original image. A region growing algorithm is applied to the read acetabulum and femur image, with seed points manually input by the user, to segment the acetabulum and femur. A femoral head identification algorithm is then applied separately to the femur to calculate the center position and maximum radius of the femoral head. The system specifically includes:

[0062] The memory stores the acquired acetabular-femoral images that require acetabular-femoral segmentation and femoral head identification, as well as the segmented acetabular position map, femoral position map, and the calculated center position and maximum radius of the femoral head.

[0063] An image acquisition device for acquiring images of the acetabulum and femur, including acetabular and femoral tissue, before image-guided hip arthroscopic minimally invasive surgery;

[0064] The processor is used in the image-guided arthroscopic minimally invasive surgery to segment the acetabular femur from preoperative images of the acetabular femur showing diseased bone tissue, and to identify the femoral head for navigation in subsequent image-guided arthroscopic minimally invasive surgery. The processor segments the acetabular femur and identifies the femoral head using the flowchart shown in Figure 1 through the following modules:

[0065] The image reading module is used to read the acetabular-femoral image that needs to be segmented and the femoral head identified, wherein the acetabular-femoral image is an image including the acetabulum and femoral tissue acquired by the image acquisition device before the image-guided hip arthroscopy minimally invasive surgery.

[0066] Specifically, the first step is to read the acetabular-femoral image of the diseased tissue used in image-guided arthroscopic minimally invasive surgery. This acetabular-femoral image can be a 3D CT image, a 2D X-ray image, or any other commonly used medical imaging method. The workflow of this invention is applicable to any type of medical imaging.

[0067] The image segmentation module is used to segment the acetabular femur image based on the range of pixel values ​​using a region growing algorithm. The region enclosed by the pixels within the range of pixel values ​​is the acetabular position segmentation map or the femur position segmentation map.

[0068] Specifically, as shown in Figure 2, the processor further includes a gradient image calculation module. Before segmenting the acetabular femur using the image segmentation module, considering that some acetabular femur images may have narrow gaps between the femur and acetabulum, or that the pixel values ​​of non-target areas and target areas are not significantly different, a gradient algorithm is defined to distinguish the femoral edge, acetabular edge, and non-target areas. The user can choose whether to enable the gradient algorithm based on the actual situation of the acetabular femur image. The gradient algorithm calculates the gradient of the acetabular femur image, obtains the gradient image of the acetabular femur image, and superimposes the gradient image onto the original acetabular femur image. The gradient algorithm specifically includes:

[0069] The gradient image is obtained using a Gaussian function for image filtering and edge detection. The Gaussian function parameter σ is selected based on the data of the acetabular femur image, including image noise level, image resolution, image contrast and brightness, and the size of the target region. A higher image noise level corresponds to a larger Gaussian function parameter σ, as do lower image resolutions, lower image contrast and brightness, and a smaller target region.

[0070] The Gaussian function parameter σ is substituted into the Gaussian function to calculate the first-order partial derivative of the Gaussian function, thereby obtaining the first-order gradient function of the acetabular femur image. The first-order gradient function is discretized to obtain a first-order gradient operator. This first-order gradient operator is then convolved with the acetabular femur image, and the gradient magnitude of each pixel is calculated to obtain the gradient image representing the image gradient magnitude. The Gaussian function parameter σ is selected by the user based on the actual situation of the acetabular femur image. Figure 3 shows the gradient image of the acetabular femur image when σ = 5.

[0071] The principle behind using the Gaussian function to obtain gradient images is as follows: The Gaussian function used in this invention is a commonly used technique in image filtering and edge detection. Gaussian filtering is a typical low-pass filter that smooths images. Furthermore, the first and second derivatives of the Gaussian function can also be used for high-pass filtering; for example, the first derivative of the Gaussian function is used in the Canny operator, and the second derivative is used in the LOG operator. The definitions of the one-dimensional and two-dimensional Gaussian functions, as well as the formulas for their first and second derivatives, are outlined below.

[0072] One-dimensional Gaussian function:

[0073] Two-dimensional Gaussian function:

[0074] The first partial derivative of the second-order Gaussian function is:

[0075] The second-order partial derivative of the two-dimensional Gaussian function is:

[0076] In image processing, the first and second gradients of a two-dimensional Gaussian function are defined as follows:

[0077] The directional gradient is (angle θ in radians):

[0078] Discretizing the gradient function yields first-order and second-order gradient operators. Convolving these operators with the image allows us to obtain the first-order and second-order gradients of the image, as well as gradients in various directions.

[0079] As shown in Figure 4, after determining whether to use the gradient algorithm to process the image, the region growing algorithm is formally used to segment the acetabular-femur image, obtaining the acetabular position segmentation map and the femur position segmentation map, specifically:

[0080] Click n+1 pixels on the acetabular femur image as seed points. The first n seed points are on the target region, and the (n+1)th seed point is on the non-target region.

[0081] The upper and lower bound thresholds for segmenting the acetabular femur image are calculated based on the pixel values ​​of the clicked n+1 seed points, specifically as follows:

[0082] Calculate the average pixel value of the first n seed points, and use it as mid_val, i.e.

[0083] Then, based on the characteristics of bone density, the average pixel value of the first n seed points is added to a first preset value to serve as the upper threshold. In this embodiment, the first preset value is taken as 350, that is, the upper threshold is upper_thresh = mid_val + 350.

[0084] The average pixel value of the first n seed points is subtracted from the second preset value to obtain the lower bound preset value. In this embodiment, the second preset value is taken as 350, that is, the lower bound preset value is mid_val-350. The lower bound preset value is then compared with the pixel value of the (n+1)th seed point (black_val = input). n The lower threshold is determined by comparing the lower threshold preset value with the pixel value of the (n+1)th seed point. That is, the lower threshold is lower_thresh = max(mid_val - 350, black_val).

[0085] Starting from the seed point, a breadth-first search is performed on the acetabular-femoral image, using the upper and lower bound thresholds as boundaries. Breadth-first search is an algorithm that traverses a graph storage structure, visiting vertices in the graph to ensure each vertex is visited only once. For the current pixel, each vertex in the acetabular-femoral image is traversed sequentially in six directions: top, bottom, left, right, front, and back. All pixels that meet the conditions are set to 1, and those that do not meet the conditions are set to 0. When the iteration stops, all pixels that satisfy the upper and lower bound thresholds are obtained as the target segmentation result. The image corresponding to the target segmentation result is either the acetabular position segmentation map or the femoral position segmentation map. Figure 5 shows the femoral position segmentation map, and Figure 6 shows the acetabular position segmentation map. Since the femur and acetabulum in the image have obvious gaps, either naturally or after gradient processing, they are easily segmented.

[0086] Furthermore, the image segmentation module further includes: smoothing the edges of the target segmentation result, specifically:

[0087] The edge regions of the target segmentation result are obtained using edge detection algorithms including Sobel and Canny.

[0088] The acquired edge region is dilated to expand the edge region by a preset number of pixels, ensuring that subsequent smoothing operations do not lose the edge information of the target segmentation result.

[0089] The expanded edge region is smoothed by applying smoothing filters, including Gaussian filtering and bilateral filtering, to reduce edge noise and discontinuities, thus making the edges smooth.

[0090] An erosion operation is performed on the smoothed edge region to restore the edge region to its original size, retain the main information of the edge, and remove unnecessary pixels introduced during the smoothing process.

[0091] The smoothed edge region is merged with the original target segmentation result to obtain the final smoothed target segmentation result.

[0092] The femoral head recognition module, as shown in Figure 7, is used to identify the femoral head based on the femoral position segmentation map using a femoral head recognition algorithm, and to calculate the center position and maximum radius of the femoral head. Specifically:

[0093] (1) Calculate the position of the center of the sphere, specifically as follows:

[0094] In the image segmentation module, the pixel values ​​of the pixels in the segmented femoral position segmentation image are marked with 1 and 0, where 1 represents the femoral bone and 0 represents the non-target region.

[0095] As shown in Figures 7(a) and (b), slice data is taken from the positive z-axis direction towards the origin. The pixel values ​​in the slice data are summed. When the first slice data where the sum of the pixel values ​​is not 0 is encountered, the average x-axis coordinate values ​​of the non-zero pixels in the slice data are taken as the x-coordinate of the center of the sphere, denoted as x. The average y-axis coordinate values ​​of the non-zero pixels in the slice data are taken as the y-coordinate of the center of the sphere, denoted as y. At the same time, the z-axis coordinate of the slice data at this time is denoted as z1.

[0096] As shown in Figures 7(c) and (d), slice data is taken from the positive x-axis direction towards the origin. The pixel values ​​in the slice data are summed. When the first slice data where the sum of the pixel values ​​is not 0 is encountered, the average value of the z-axis coordinate of the non-zero pixel points in the slice data is the z-coordinate of the center of the sphere, denoted as z. At the same time, the x-axis coordinate of the slice data at this time is denoted as x1.

[0097] (2) Calculate the maximum radius as follows:

[0098] Find the maximum value between z1-z and x1-x, where the maximum value is the maximum radius. That is, r = max(z1-z, x1-x). Figure 8 is a cross-sectional view drawn based on the calculated center of the femoral head and radius.

[0099] Second Embodiment

[0100] As shown in Figure 9, this embodiment provides a method for segmenting the acetabular femur and identifying the femoral head using a system for segmenting the acetabular femur and identifying the femoral head as described in the first embodiment, including:

[0101] S1: Read the acetabular-femoral image that needs to be segmented and the femoral head identified, wherein the acetabular-femoral image is an image of the acetabulum and femoral tissue acquired by medical imaging equipment before the image-guided hip arthroscopy minimally invasive surgery.

[0102] S2: Using a region growing algorithm, the acetabular femur image is segmented according to the range of pixel values. The region enclosed by the pixels within the range of pixel values ​​is used to obtain the acetabular position segmentation map or the femur position segmentation map.

[0103] S3: Based on the femoral position segmentation map, a femoral head recognition algorithm is used to identify the femoral head and calculate the center position and maximum radius of the femoral head.

[0104] A computer-readable storage medium stores computer code that, when executed, performs the methods described above. Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. This program can be stored in a computer-readable storage medium, which may include: read-only memory (ROM), random access memory (RAM), a magnetic disk, or an optical disk, etc.

[0105] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should also be considered within the scope of protection of the present invention.

[0106] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0107] It should be noted that the above embodiments can be freely combined as needed. The above description is only a preferred embodiment of the present invention. It should be pointed out that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A system for segmenting an acetabulum femur and identifying a femoral head, characterized by, The application relates to an image guided hip arthroscopy system, which comprises the following parts: a memory for storing the acquired acetabulum-femur image which needs to be segmented and the femoral head to be identified, and the segmented acetabular position segmentation image, femoral position segmentation image, calculated spherical center position and maximum radius of the femoral head; an image acquisition device for acquiring the acetabulum-femur image including acetabular and femoral tissue before the image guided hip arthroscopy surgery; a processor for segmenting the acetabulum-femur in the image guided hip arthroscopy surgery and identifying the femoral head in the acetabulum-femur image acquired before the surgery which presents pathological bone tissue, so as to be navigated in the subsequent image guided hip arthroscopy surgery and segmented and identified by the following modules: an image reading module for reading the acetabulum-femur image which needs to be segmented and the femoral head to be identified, wherein the acetabulum-femur image is the image including acetabular and femoral tissue acquired by the image acquisition device before the image guided hip arthroscopy surgery; an image segmentation module for segmenting the acetabulum-femur image according to the range of pixel value of the pixel point by using a region growing algorithm, wherein the region surrounded by the pixel point in the range of pixel value is the acetabular position segmentation image or femoral position segmentation image; a femoral head identification module for identifying the femoral head by using a femoral head identification algorithm based on the femoral position segmentation image, and calculating the spherical center position and maximum radius of the femoral head.

2. The system for dividing the acetabulum femur and identifying the femoral head according to claim 1, characterized in that, The processor further comprises a gradient image calculation module; the gradient image calculation module is used for calculating the gradient of the acetabulum-femur image, acquiring the gradient image of the acetabulum-femur image and superimposing the gradient image on the original acetabulum-femur image when the gap between the femur and the acetabulum in the read acetabulum-femur image is narrow or the pixel value difference between the non-target region and the target region is not obvious before the acetabulum-femur is segmented by using the image segmentation module, wherein the target region is the region where the acetabulum or the femur is located, and the non-target region is the region other than the region where the acetabulum or the femur is located.

3. The system for dividing the acetabulum femur and identifying the femoral head according to claim 2, characterized in that, In the gradient image calculation module, the gradient of the acetabulum-femur image is calculated, the gradient image of the acetabulum-femur image is acquired, and the gradient image is superimposed on the original acetabulum-femur image, specifically as follows: the gradient image is acquired by using a Gaussian function for image filtering and edge detection; the Gaussian function parameter sigma is selected, wherein the Gaussian function parameter sigma is selected according to the data of the acetabulum-femur image including image noise level, image resolution, image contrast and brightness, and the size of the target region, the higher the image noise level, the larger the Gaussian function parameter sigma, the lower the image resolution, the larger the Gaussian function parameter sigma, the lower the image contrast and brightness, the larger the Gaussian function parameter sigma, and the smaller the target region, the larger the Gaussian function parameter sigma; the first-order partial derivative of the Gaussian function is calculated by substituting the Gaussian function parameter sigma into the Gaussian function, so as to obtain the first-order gradient function of the acetabulum-femur image. Discretize the first-order gradient function to obtain a first-order gradient operator, and perform convolution of the first-order gradient operator and the acetabulum-femur image to calculate the gradient size of each pixel point respectively, and obtain the gradient image representing the gradient size of the image.

4. The system for dividing the acetabulum femur and identifying the femoral head of claim 2, wherein, In the image segmentation module, the region growing algorithm is used to segment the acetabulum-femur image according to the range of pixel values of the pixel points, and the region surrounded by the pixel points in the range of pixel values is the acetabulum position segmentation graph or the femur position segmentation graph, specifically: Click n+1 pixel points on the acetabulum-femur image as seed points, the first n seed points are on the target region, and the n+1 seed point is on the non-target region; Calculate the upper threshold and the lower threshold for segmenting the acetabulum-femur image through the pixel values of the n+1 clicked seed points; Starting from the seed points, the upper threshold and the lower threshold are used as boundaries to perform breadth-first search on the acetabulum-femur image to obtain a target segmentation result, wherein the breadth-first search is an algorithm for traversing image storage structure, each vertex in the acetabulum-femur image is traversed from the six directions of up, down, left, right, front and back one by one, and all pixel points satisfying the upper threshold and the lower threshold are obtained as the target segmentation result, and the image corresponding to the target segmentation result is the acetabulum position segmentation graph or the femur position segmentation graph.

5. The system for dividing the acetabulum femur and identifying the femoral head of claim 4, wherein, In the image segmentation module, the upper threshold and the lower threshold for segmenting the acetabulum-femur image are calculated through the pixel values of the n+1 clicked seed points, specifically: Calculate the mean value of the pixel values of the first n seed points; Add a first preset value to the mean value of the pixel values of the first n seed points as the upper threshold; Subtract a second preset value from the mean value of the pixel values of the first n seed points as a lower preset value, and compare the lower preset value with the pixel value of the n+1 seed point, and take the larger value of the lower preset value and the pixel value of the n+1 seed point as the lower threshold.

6. The system for dividing the acetabulum femur and identifying the femoral head of claim 4, wherein, The image segmentation module further includes smoothing the edges of the target segmentation result, specifically: Obtain the edge region of the target segmentation result through an edge detection algorithm including Sobel and Canny; Perform an inflation operation on the obtained edge region to expand the edge region by a preset pixel, so as to ensure that the subsequent smoothing operation does not lose the edge information of the target segmentation result; Apply a smoothing filter including Gaussian filter and bilateral filter to the expanded edge region to smooth the edge, so as to reduce the noise and discontinuity of the edge and make the edge smooth; Perform an erosion operation on the smoothed edge region to restore the edge region to the original size, retain the main information of the edge, and remove unnecessary pixels introduced in the smoothing process; Merge the smoothed edge region with the original target segmentation result to obtain the final smoothed target segmentation result.

7. The system for dividing the acetabulum femur and identifying the femoral head of claim 4, wherein, In the femoral head recognition module, the ball center position is calculated, specifically: In the image segmentation module, the pixel value of the pixel point in the segmented femur position segmentation image is marked with 1 and 0, 1 is the femur, and 0 is the non-target region; The slice data is taken from the positive direction of the z-axis to the origin, the sum of the pixel values in the slice data is calculated, the first slice data with a non-zero pixel value sum is encountered, the average value of the x-axis coordinate value of the non-zero pixel point in the slice data is the x-coordinate of the ball center position, denoted as x, the average value of the y-axis coordinate value of the non-zero pixel point in the slice data is the y-coordinate of the ball center position, denoted as y, and the z-axis coordinate of the slice data at this time is denoted as z1; The slice data is taken from the positive direction of the x-axis to the origin, the sum of the pixel values in the slice data is calculated, the first slice data with a non-zero pixel value sum is encountered, the average value of the z-axis coordinate value of the non-zero pixel point in the slice data is the z-coordinate of the ball center position, denoted as z, and the x-axis coordinate of the slice data at this time is denoted as x1.

8. The system for dividing the acetabulum femur and identifying the femoral head according to claim 7, characterized in that, In the image segmentation module, the maximum radius is calculated, specifically: The maximum value of z1-z and x1-x is obtained, and the maximum value is the maximum radius.

9. A method of segmenting a hip joint using a system according to any one of claims 1 to 8, wherein the method comprises: - segmenting the hip joint using the system; and - identifying the femoral head using the system. Comprise: S1: reading a hip femur image which needs to be segmented and the femoral head of the hip femur image is identified, wherein the hip femur image is an image including acetabulum and femur tissue obtained by a medical imaging device before the image guided hip arthroscopy minimally invasive surgery; S2: using a region growing algorithm, the acetabulum femur image is segmented according to the range of pixel value of the pixel point, and the region surrounded by the pixel point in the range of pixel value is the acetabulum position segmentation image or the femur position segmentation image; S3: based on the femur position segmentation image, a femur recognition algorithm is used to identify the femoral head, and the ball center position and the maximum radius of the femoral head are calculated.

10. A computer readable storage medium, the computer readable storage medium stores computer code, when the computer code is executed, the method as claimed in claim 9 is executed.

Images