Bone image processing method, device, equipment and medium

Abstract

The present disclosure relates to a method, device, equipment and medium for processing bone images, the method comprising: obtaining an original segmentation mask of a target bone, an original bone surface model, a mirror segmentation mask corresponding to the original segmentation mask, and a mirror bone surface model corresponding to the original bone surface model; registering the original bone surface model with the mirror bone surface model to obtain a transformation matrix; registering the original segmentation mask and the mirror segmentation mask according to the transformation matrix; performing Boolean value operation on the registered original segmentation mask and the mirror segmentation mask to determine a first osteophyte region; removing the first osteophyte region in the original segmentation mask to obtain a candidate segmentation mask; determining a second osteophyte region based on the candidate segmentation mask and a preset segmentation mask template; removing the second osteophyte region in the candidate segmentation mask to obtain a target segmentation mask. The present disclosure can improve the accuracy of osteophyte removal and more accurately restore the original bone structure and surface of the target bone.

Description

Technical Field

[0001] The present disclosure relates to the technical field of medical image processing, and in particular to a method, device, equipment and medium for processing bone images. Background Art

[0002] In the treatment of joints such as the knee, hip and ankle, the structural analysis of the joint bone tissue and the calculation of parameters such as the axis are particularly important, which directly determine the model and placement of the prosthesis, the grinding position of the bone, and then determine the patient's postoperative recovery and treatment effect. In order to accurately perform structural analysis and parameter calculation, it is necessary to first realize bone image segmentation. However, the patient's joint position is often accompanied by abnormal proliferation of bone structure, namely osteophytes, also known as bone spurs. This abnormal proliferation structure will change the original bone joint surface structure and characteristics, and will have a very adverse effect on the model selection and location planning of the prosthesis. Therefore, how to accurately identify osteophytes and restore the original bone structure and surface is a major technical difficulty in bone image processing. Summary of the invention

[0003] In order to solve the above technical problems, the present disclosure provides a bone image processing method, device, equipment and medium.

[0004] According to one aspect of the present disclosure, a method for processing a bone image is provided, comprising:

[0005] Acquire an original segmentation mask of the target bone, an original bone surface model, a mirror image segmentation mask corresponding to the original segmentation mask, and a mirror image bone surface model corresponding to the original bone surface model;

[0006] The original bone surface model is registered with the mirror image bone surface model to obtain a transformation matrix;

[0007] aligning the original segmentation mask and the mirrored segmentation mask according to the transformation matrix;

[0008] Performing Boolean value operation on the original segmentation mask and the mirror segmentation mask after registration to determine a first osteophyte region;

[0009] removing the first osteophyte region in the original segmentation mask to obtain a candidate segmentation mask;

[0010] Determining a second osteophyte region based on the candidate segmentation mask and a preset segmentation mask template;

[0011] The second osteophyte region in the candidate segmentation mask is removed to obtain a target segmentation mask.

[0012] According to another aspect of the present disclosure, there is also provided a bone image processing device, comprising:

[0013] A bone image acquisition module, used to acquire an original segmentation mask of a target bone, an original bone surface model, and a mirror image segmentation mask corresponding to the original segmentation mask, and a mirror image bone surface model corresponding to the original bone surface model;

[0014] A first configuration module is used to register the original bone surface model with the mirror image bone surface model to obtain a transformation matrix;

[0015] A second configuration module, configured to align the original segmentation mask and the mirror segmentation mask according to the transformation matrix;

[0016] A first osteophyte determination module, configured to perform a Boolean operation on the original segmentation mask and the mirror segmentation mask after registration to determine a first osteophyte region;

[0017] A first osteophyte removal module, configured to remove the first osteophyte region in the original segmentation mask to obtain a candidate segmentation mask;

[0018] A second osteophyte determination module, configured to determine a second osteophyte region based on the candidate segmentation mask and a preset segmentation mask template;

[0019] The second osteophyte removal module is used to remove the second osteophyte region in the candidate segmentation mask to obtain a target segmentation mask.

[0020] According to another aspect of the present disclosure, an electronic device is provided, the electronic device comprising:

[0021] processor;

[0022] a memory for storing instructions executable by the processor;

[0023] The processor is used to read the executable instructions from the memory and execute the instructions to implement the above method.

[0024] According to another aspect of the present disclosure, a computer-readable storage medium is provided, wherein the storage medium stores a computer program, and the computer program is used to execute the above method.

[0025] Compared with the prior art, the technical solution provided by the embodiments of the present disclosure has the following advantages:

[0026] The bone image processing method provided by the embodiment of the present disclosure includes: obtaining the original segmentation mask of the target bone, the original bone surface model, the mirror segmentation mask corresponding to the original segmentation mask, and the mirror bone surface model corresponding to the original bone surface model. The original bone surface model and the mirror bone surface model are aligned to obtain a transformation matrix; since the original bone surface model and the mirror bone surface model show bone surface data, the contour area of ​​the target bone can be accurately reflected, and the alignment can be accurately achieved. Moreover, since it is only surface data, only a small number of key points are used for alignment, which improves the alignment efficiency. Secondly, the original segmentation mask and the mirror segmentation mask are aligned according to the transformation matrix; the alignment efficiency and alignment accuracy between the two segmentation masks can be improved. The original segmentation mask and the mirror segmentation mask after registration are subjected to Boolean value operation to determine the first osteophyte area; the first osteophyte area in the original segmentation mask is removed to obtain a candidate segmentation mask; the second osteophyte area is determined based on the candidate segmentation mask and the preset segmentation mask template; the second osteophyte area in the candidate segmentation mask is removed to obtain a target segmentation mask; this method first removes the first osteophyte area in the original segmentation mask, and then further removes the second osteophyte area in combination with a healthy segmentation mask template without osteophytes. Through two Boolean value operations and removal of the osteophyte area, the osteophyte structure attached to the target bone can be more thoroughly and accurately identified and removed, the accuracy of osteophyte removal can be improved, and the original bone structure and surface of the target bone can be more accurately restored. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the present disclosure.

[0028] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0029] Figure 1 A flowchart of a method for processing bone images provided in an embodiment of the present disclosure;

[0030] Figure 2 A schematic diagram of a CT image provided by an embodiment of the present disclosure;

[0031] Figure 3 A schematic diagram of a segmentation mask provided in an embodiment of the present disclosure;

[0032] Figure 4 A structural block diagram of a bone image processing device provided in an embodiment of the present disclosure;

[0033] Figure 5 A schematic diagram of the structure of an electronic device provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0034] In order to more clearly understand the above-mentioned objectives, features and advantages of the present disclosure, the scheme of the present disclosure will be further described below. It should be noted that the embodiments of the present disclosure and the features in the embodiments can be combined with each other without conflict.

[0035] In the following description, many specific details are set forth to facilitate a full understanding of the present disclosure, but the present disclosure may also be implemented in other ways different from those described herein; it is obvious that the embodiments in the specification are only part of the embodiments of the present disclosure, rather than all of the embodiments.

[0036] In traditional technology, bone image segmentation methods or methods based on neural network deep learning can be used to segment bone images. Among them, bone image segmentation is a semi-automatic segmentation, which requires manual selection of seed points and the segmentation of bone images through region growing and cavity filling. However, traditional bone image segmentation methods have problems such as low automation, poor stability and low accuracy. Methods based on neural network deep learning often require a large number of high-quality annotated images as training data. Such data are difficult to obtain and difficult to annotate. Therefore, it is technically difficult to use this method alone to directly distinguish between osteophytes and native bone structures.

[0037] Based on this, in order to accurately identify and remove osteophytes and restore the original bone structure and surface, the embodiments of the present disclosure provide a bone image processing method, device, equipment and medium. For ease of understanding, the embodiments of the present disclosure are described in detail below.

[0038] Figure 1 The flowchart of a bone image processing method provided by the embodiment of the present disclosure can be executed by a bone image processing device configured in a terminal, and the device can be implemented by software and / or hardware. Figure 1 As shown, the bone image processing method may include the following steps:

[0039] S102, obtaining an original segmentation mask of the target bone, an original bone surface model, and a mirror image segmentation mask corresponding to the original segmentation mask and a mirror image bone surface model corresponding to the original bone surface model.

[0040] The target bone may be any bone that requires surgical treatment to restore joint function, including but not limited to: femur or tibia. The target bone has an abnormal proliferation of bone structure, namely, osteophyte structure, at the joint position.

[0041] In this embodiment, obtaining the original segmentation mask and the original bone surface model of the target bone can include: obtaining a CT (Computed Tomography) image corresponding to the target bone with an osteophyte structure; segmenting the area containing the target bone in the CT image through a deep learning model to obtain the original segmentation mask; and performing three-dimensional reconstruction on the original segmentation mask to obtain the original bone surface model.

[0042] As a specific example, in the preoperative stage, an X-ray beam may be used to perform a tomographic scan of a patient's target bone part to collect three-dimensional CT thin-layer volume data to obtain a CT image; Figure 2 The example in shows bilateral CT images of a patient's lower limbs.

[0043] In other examples, CT images can be directly retrieved from a local storage area; manually uploaded CT images can be received through a designated port; or, in some Internet medical systems, CT images can be downloaded from a server through network communication. The CT image can include target bones with osteophyte structures, such as femurs or tibias.

[0044] In this embodiment, deep learning models such as U-Net and R-CNN (Region-Convolutional Neural Networks) can be used to segment the region containing the target bone in the CT image to obtain the original segmentation mask. Figure 3 The original segmentation mask is shown, where osteophyte structures are clearly visible at the joint location.

[0045] The bone surface of the original segmentation mask is reconstructed in three dimensions to obtain the original bone surface model. The original bone surface model can also be understood as the bone surface extracted from the original segmentation mask; relative to the original segmentation mask, the original bone surface model is a hollow model that only displays the bone surface. Since the osteophyte structure is attached to the bone surface, the original bone surface model can clearly and completely display the surface features of the target bone and osteophyte structure.

[0046] In this embodiment, obtaining a mirror segmentation mask corresponding to the original segmentation mask and a mirror bone surface model corresponding to the original bone surface model may include: mirroring the original segmentation mask to obtain a mirror segmentation mask; mirroring the original bone surface model, or three-dimensionally reconstructing the mirror segmentation mask to obtain a mirror bone surface model.

[0047] In a specific example, the original segmentation mask of the operated leg can be mirrored left and right to obtain a corresponding mirror segmentation mask. Similarly, the original bone surface model can be mirrored left and right to obtain a corresponding mirror bone surface model. Alternatively, in another way, the mirror segmentation mask can be reconstructed in three dimensions to obtain a mirror bone surface model.

[0048] S104, registering the original bone surface model with the mirror image bone surface model to obtain a transformation matrix.

[0049] In this embodiment, the original bone surface model and the mirror image bone surface model are registered to obtain a transformation matrix, which can refer to the following specific steps (1) to (4):

[0050] (1) Select multiple first key points at the first position of the original bone surface model. Taking the femur as an example of the target bone, the first position may include: a bone surface determined based on the medial and lateral condyles of the femur. In this embodiment, N (e.g., N=16) first key points may be selected on the bone surface of the non-osteophyte area near the medial and lateral condyles of the femur of the original bone surface model.

[0051] (2) Selecting a plurality of second key points at a second position of the mirrored bone surface model; wherein the first position and the second position correspond to the same position of the target bone, and following the above example, the second position can also be a bone surface determined based on the medial and lateral condyles of the femur. In this embodiment, N second key points can be selected on the bone surface of the non-osteophyte area near the medial and lateral condyles of the femur of the mirrored bone surface model.

[0052] (3) based on the first key point and the second key point, the original bone surface model and the mirror image bone surface model are carried out to coarse registration, and registration position information is obtained. In the specific implementation process, based on the first key point and the second key point, and using coarse registration algorithms such as least square method or Kabsch method to solve the rigid transformation matrix, the coarse registration of the original bone surface model and the mirror image bone surface model is realized, and registration position information is obtained. This registration position information can for example include horizontal offset, vertical offset and rotation angle.

[0053] (4) Based on the registration position information, the original bone surface model and the mirror image bone surface model are precisely registered to obtain a transformation matrix.

[0054] This embodiment provides registration position information according to the rough registration result, and uses the registration position information for the fine registration process. Specifically, the original bone surface model and the mirrored bone surface model are finely registered according to the registration position information and a preset fine registration algorithm is used to obtain their transformation matrix; wherein, the preset fine registration algorithm is, for example, ICP (Iterative Closest Point) registration algorithm, RANSAC (Random Sample Consensus), ProSAC (Progressive Sampling with Approximate Clustering) and configuration algorithm based on deep learning, etc.

[0055] This embodiment can improve the registration accuracy between the original bone surface model and the mirror image bone surface model and improve the accuracy of the transformation matrix through the two-stage registration of coarse registration and fine registration.

[0056] S106, registering the original segmentation mask and the mirror image segmentation mask according to the transformation matrix.

[0057] Specifically, the coordinates of each pixel in the original segmentation mask can be substituted into the above transformation matrix for coordinate transformation to obtain the original segmentation mask after being aligned with the mirror segmentation mask; or, the coordinates of each pixel in the mirror segmentation mask can be substituted into the above transformation matrix for coordinate transformation to obtain the mirror segmentation mask after being aligned with the original segmentation mask.

[0058] In the above two segmentation mask registration schemes of this embodiment, the original bone surface model representing its bone surface features is first registered with the mirror image bone surface model; since the above bone surface model only has bone surface data, it can be registered with only a small number of key points, which can improve the registration efficiency; at the same time, the registration accuracy is improved through coarse registration and fine registration during the registration process. Therefore, the original segmentation mask and the mirror image segmentation mask are registered using the transformation matrix after the original bone surface model and the mirror image bone surface model are registered, which can effectively improve the registration efficiency and registration accuracy.

[0059] S108, performing Boolean value operation on the original segmentation mask after registration and the mirror image segmentation mask to determine a first osteophyte region.

[0060] In this embodiment, a Boolean value operation is performed on the original segmentation mask and the mirror segmentation mask after registration, and the difference between the mask union and the mask intersection of the two is determined as the first osteophyte region containing the osteophyte structure.

[0061] S110, removing the first osteophyte region in the original segmentation mask to obtain a candidate segmentation mask.

[0062] S112: Determine a second osteophyte region based on the candidate segmentation mask and a preset segmentation mask template.

[0063] In order to remove osteophytes more accurately, on the basis of removing the first osteophyte region, this embodiment may combine the segmentation mask template to further remove the residual osteophyte structure that has not been completely removed on the candidate segmentation mask.

[0064] The segmentation mask template includes a healthy sample bone without osteophytes, and the sample bone is the same bone as the target bone to be treated, such as the femur or tibia. For ease of understanding, this embodiment provides a method for obtaining a segmentation mask template, as shown below.

[0065] Acquire multiple sample CT images including sample bones; wherein the sample bones are free of osteophyte structures; segment the regions including the sample bones in the multiple sample CT images through a deep learning model to obtain multiple sample segmentation masks; annotate the bone structure parameters for the multiple sample segmentation masks; based on the bone structure parameters and a preset parameter interval, select a portion of the multiple sample segmentation masks as a segmentation mask template.

[0066] In this embodiment, the bone structure parameters annotated for each sample segmentation mask include, for example, centerline length, anterior and posterior condyle length, and medial and lateral condyle length, etc. Then, a statistical rule is determined based on the bone structure parameters, and the specific statistical rule is, for example, a rule of sorting from long to short centerline length, a rule of sorting from long to short anterior and posterior condyle length, and / or a rule of sorting from long to short medial and lateral condyle length.

[0067] According to the above statistical rules, multiple sample segmentation masks are statistically sorted. Among the multiple sample segmentation masks arranged in order, a part is selected according to the preset parameter interval, and the parameter interval is, for example, one sample segmentation mask is selected every 10, one sample segmentation mask is selected every 1 cm of the centerline length, one sample segmentation mask is selected every 1 cm of the anterior and posterior condyle length, etc. In this way, multiple sample segmentation masks with different parameters and representativeness can be selected, and the selected part of the sample segmentation masks is determined as the segmentation mask template.

[0068] On the basis of the above embodiments, determining the second osteophyte region based on the candidate segmentation mask and the preset segmentation mask template may include the following embodiments.

[0069] (1) Aligning the candidate segmentation mask with multiple preset segmentation mask templates. The specific implementation of this step can refer to the embodiment of aligning the original segmentation mask and the mirror segmentation mask shown in the above steps S104-S106, and will not be described in detail here.

[0070] (2) Calculate the DICE coefficients between the candidate segmentation mask after registration and the multiple segmentation mask templates respectively. The DICE coefficient is used to measure the similarity between the candidate segmentation mask after registration and each segmentation mask template. In this embodiment, the DICE coefficient is used as an important indicator for evaluating the registration effect between the candidate segmentation mask and the segmentation mask template. It can be understood that the higher the DICE coefficient, the better the registration effect between the candidate segmentation mask and the segmentation mask template.

[0071] (3) Determine the target segmentation mask template with the largest DICE coefficient between it and the candidate segmentation mask from multiple segmentation mask templates. The target segmentation mask template with the largest DICE coefficient has the best registration effect with the candidate segmentation mask, and the sample bone contained in it is closest to the target bone in a healthy state after osteophyte removal.

[0072] (4) Performing a Boolean operation on the candidate segmentation mask and the target segmentation mask template to determine the second osteophyte region.

[0073] Specifically, a Boolean value operation is performed on the registered candidate segmentation mask and the target segmentation mask template, and the difference between the mask union and the mask intersection of the two is determined as the second osteophyte region containing the osteophyte structure.

[0074] S114, removing the second osteophyte region in the candidate segmentation mask to obtain a target segmentation mask.

[0075] In this embodiment, the second osteophyte region is further removed from the candidate segmentation mask to obtain a target segmentation mask after more thorough osteophyte removal. Figure 3 On the right, the target segmentation mask is shown. Compared with the original segmentation mask, the osteophyte structure at the joint has been clearly removed more thoroughly.

[0076] This embodiment can also perform three-dimensional reconstruction on the target segmentation mask to obtain a target bone surface model. In practical applications, the target bone surface model can be used to make a prosthesis.

[0077] In summary, the bone image processing method provided by the embodiment of the present disclosure includes: obtaining the original segmentation mask, the original bone surface model, the mirror segmentation mask corresponding to the original segmentation mask, and the mirror bone surface model corresponding to the original bone surface model of the target bone; aligning the original bone surface model with the mirror bone surface model to obtain a transformation matrix; aligning the original segmentation mask and the mirror segmentation mask according to the transformation matrix; performing Boolean value operation on the aligned original segmentation mask and the mirror segmentation mask to determine the first osteophyte area; removing the first osteophyte area in the original segmentation mask to obtain a candidate segmentation mask; determining the second osteophyte area based on the candidate segmentation mask and a preset segmentation mask template; removing the second osteophyte area in the candidate segmentation mask to obtain a target segmentation mask.

[0078] In the technical scheme, the original bone surface model and the mirrored bone surface model, which are easy to register, are first used to register to obtain a transformation matrix; since the original bone surface model and the mirrored bone surface model show bone surface data, the contour area of ​​the target bone can be accurately reflected, and the registration can be accurately achieved. Moreover, since it is only surface data, only a small number of key points can be used for registration, which improves the registration efficiency. Secondly, the original segmentation mask and the mirrored segmentation mask are registered according to the transformation matrix, which can improve the registration efficiency and registration accuracy between the two segmentation masks. The first osteophyte area in the original segmentation mask is determined by Boolean value operation. After the first osteophyte area is preliminarily removed, the second osteophyte area is further removed in combination with the healthy segmentation mask template without osteophytes. By performing two Boolean value operations and removing the osteophyte area, the osteophyte structure attached to the target bone can be more thoroughly and accurately identified and removed, the accuracy of osteophyte removal is improved, and the original bone structure and surface of the target bone are more accurately restored.

[0079] Figure 4 FIG. 1 is a structural block diagram of a bone image processing device provided in an embodiment of the present disclosure, and the device can be implemented by software and / or hardware. Figure 4 As shown, the bone image processing device may include the following modules:

[0080] The bone image acquisition module 210 is used to acquire the original segmentation mask of the target bone, the original bone surface model, the mirror image segmentation mask corresponding to the original segmentation mask, and the mirror image bone surface model corresponding to the original bone surface model;

[0081] A first configuration module 220 is used to register the original bone surface model with the mirror image bone surface model to obtain a transformation matrix;

[0082] A second configuration module 230, configured to align the original segmentation mask and the mirror segmentation mask according to the transformation matrix;

[0083] A first osteophyte determination module 240 is used to perform a Boolean operation on the original segmentation mask and the mirror segmentation mask after registration to determine a first osteophyte region;

[0084] A first osteophyte removal module 250, configured to remove the first osteophyte region in the original segmentation mask to obtain a candidate segmentation mask;

[0085] A second osteophyte determination module 260, configured to determine a second osteophyte region based on the candidate segmentation mask and a preset segmentation mask template;

[0086] The second osteophyte removal module 270 is configured to remove the second osteophyte region in the candidate segmentation mask to obtain a target segmentation mask.

[0087] In one embodiment, the first configuration module 220 is further configured to:

[0088] Selecting a plurality of first key points at a first position of the original bone surface model;

[0089] Selecting a plurality of second key points at a second position of the mirror image bone surface model; wherein the first position and the second position correspond to the same position of the target bone;

[0090] Based on the first key point and the second key point, the original bone surface model and the mirror image bone surface model are roughly registered to obtain registration position information;

[0091] Based on the registration position information, the original bone surface model and the mirror image bone surface model are precisely registered to obtain a transformation matrix.

[0092] In one embodiment, the second osteophyte determination module 260 is further configured to:

[0093] Aligning the candidate segmentation mask with a plurality of preset segmentation mask templates;

[0094] respectively calculating DICE coefficients between the candidate segmentation mask after registration and the plurality of segmentation mask templates;

[0095] Determine, from the plurality of segmentation mask templates, a target segmentation mask template having a maximum DICE coefficient with the candidate segmentation mask;

[0096] A Boolean operation is performed on the candidate segmentation mask and the target segmentation mask template to determine a second osteophyte region.

[0097] In one embodiment, the apparatus further includes: a template determination module, which is used to:

[0098] Acquire a plurality of sample CT images including sample bones; wherein the sample bones do not have osteophyte structures attached thereto;

[0099] Segmenting the regions containing the sample bones in the plurality of sample CT images by a deep learning model to obtain a plurality of sample segmentation masks;

[0100] labeling bone structure parameters for a plurality of the sample segmentation masks;

[0101] Based on the bone structure parameters and a preset parameter interval, a portion of the plurality of sample segmentation masks is selected as a segmentation mask template.

[0102] In one embodiment, the bone image acquisition module 210 is further used for:

[0103] Acquire a CT image corresponding to the target bone with osteophyte structure;

[0104] Segmenting the region containing the target bone in the CT image by a deep learning model to obtain an original segmentation mask;

[0105] The original segmentation mask is three-dimensionally reconstructed to obtain an original bone surface model.

[0106] In one embodiment, the bone image acquisition module 210 is further used for:

[0107] Performing mirror processing on the original segmentation mask to obtain a mirror segmentation mask;

[0108] The original bone surface model is mirrored, or the mirror segmentation mask is three-dimensionally reconstructed to obtain a mirrored bone surface model.

[0109] In one embodiment, the target bone includes: a femur or a tibia.

[0110] The implementation principle and technical effects of the device provided in this embodiment are the same as those of the aforementioned method embodiment. For the sake of brief description, for matters not mentioned in the device embodiment, reference may be made to the corresponding contents in the aforementioned method embodiment.

[0111] Figure 5 The structure diagram of an electronic device provided by the embodiment of the present disclosure is shown in FIG. Figure 5 As shown, the electronic device 300 includes one or more processors 301 and a memory 302 .

[0112] The processor 301 may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device 300 to perform desired functions.

[0113] The memory 302 may include one or more computer program products, and the computer program product may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, a random access memory (RAM) and / or a cache memory (cache), etc. The non-volatile memory may include, for example, a read-only memory (ROM), a hard disk, a flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 301 may run the program instructions to implement the bone image processing method of the embodiment of the present disclosure described above and / or other desired functions. Various contents such as input signals, signal components, noise components, etc. may also be stored in the computer-readable storage medium.

[0114] In one example, the electronic device 300 may further include: an input device 303 and an output device 304 , and these components are interconnected via a bus system and / or other forms of connection mechanisms (not shown).

[0115] In addition, the input device 303 may also include, for example, a keyboard, a mouse, and the like.

[0116] The output device 304 can output various information to the outside, including the determined distance information, direction information, etc. The output device 304 can include, for example, a display, a speaker, a printer, a communication network and a remote output device connected thereto, and the like.

[0117] Of course, to simplify, Figure 5 Only some of the components related to the present disclosure in the electronic device 300 are shown, and components such as a bus, an input / output interface, etc. are omitted. In addition, according to specific application situations, the electronic device 300 may also include any other appropriate components.

[0118] Furthermore, this embodiment also provides a computer-readable storage medium, wherein the storage medium stores a computer program, and the computer program is used to execute the above-mentioned bone image processing method.

[0119] The computer program product of a bone image processing method, device, electronic device and medium provided in the embodiments of the present disclosure includes a computer-readable storage medium storing program code. The instructions included in the program code can be used to execute the method described in the previous method embodiments. The specific implementation can be found in the method embodiments and will not be repeated here.

[0120] It should be noted that, in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0121] The above description is only a specific embodiment of the present disclosure, so that those skilled in the art can understand or implement the present disclosure. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present disclosure. Therefore, the present disclosure will not be limited to the embodiments described herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for processing a bone image, characterized in that: include: Acquire an original segmentation mask of the target bone, an original bone surface model, a mirror image segmentation mask corresponding to the original segmentation mask, and a mirror image bone surface model corresponding to the original bone surface model; The original bone surface model is registered with the mirror image bone surface model to obtain a transformation matrix; aligning the original segmentation mask and the mirrored segmentation mask according to the transformation matrix; Performing Boolean value operation on the original segmentation mask and the mirror segmentation mask after registration to determine a first osteophyte region; removing the first osteophyte region in the original segmentation mask to obtain a candidate segmentation mask; Determining a second osteophyte region based on the candidate segmentation mask and a preset segmentation mask template; The second osteophyte region in the candidate segmentation mask is removed to obtain a target segmentation mask.

2. The method according to claim 1, characterized in that The step of registering the original bone surface model with the mirror image bone surface model to obtain a transformation matrix comprises: Selecting a plurality of first key points at a first position of the original bone surface model; Selecting a plurality of second key points at a second position of the mirror image bone surface model; wherein the first position and the second position correspond to the same position of the target bone; Based on the first key point and the second key point, the original bone surface model and the mirror image bone surface model are roughly registered to obtain registration position information; Based on the registration position information, the original bone surface model and the mirror image bone surface model are precisely registered to obtain a transformation matrix.

3. The method according to claim 1, characterized in that The determining the second osteophyte region based on the candidate segmentation mask and a preset segmentation mask template includes: Aligning the candidate segmentation mask with a plurality of preset segmentation mask templates; respectively calculating DICE coefficients between the candidate segmentation mask after registration and the plurality of segmentation mask templates; Determine, from the plurality of segmentation mask templates, a target segmentation mask template having a maximum DICE coefficient with the candidate segmentation mask; A Boolean operation is performed on the candidate segmentation mask and the target segmentation mask template to determine a second osteophyte region.

4. The method according to claim 1, characterized in that The method further comprises: Acquire a plurality of sample CT images including sample bones; wherein the sample bones do not have osteophyte structures attached thereto; Segmenting the regions containing the sample bones in the plurality of sample CT images by a deep learning model to obtain a plurality of sample segmentation masks; labeling bone structure parameters for a plurality of the sample segmentation masks; Based on the bone structure parameters and a preset parameter interval, a portion of the plurality of sample segmentation masks is selected as a segmentation mask template.

5. The method according to claim 1, characterized in that The method of obtaining the original segmentation mask and the original bone surface model of the target bone includes: Acquire a CT image corresponding to the target bone with osteophyte structure; Segmenting the region containing the target bone in the CT image by a deep learning model to obtain an original segmentation mask; The original segmentation mask is three-dimensionally reconstructed to obtain an original bone surface model.

6. The method according to claim 1, characterized in that Acquiring a mirror image segmentation mask corresponding to the original segmentation mask and a mirror image bone surface model corresponding to the original bone surface model, comprising: Performing mirror processing on the original segmentation mask to obtain a mirror segmentation mask; The original bone surface model is mirrored, or the mirror segmentation mask is three-dimensionally reconstructed to obtain a mirrored bone surface model.

7. The method according to claim 1, characterized in that The target bone includes: a femur or a tibia.

8. A bone image processing device, characterized in that: include: A bone image acquisition module, used to acquire an original segmentation mask of a target bone, an original bone surface model, and a mirror image segmentation mask corresponding to the original segmentation mask, and a mirror image bone surface model corresponding to the original bone surface model; A first configuration module is used to register the original bone surface model with the mirror image bone surface model to obtain a transformation matrix; A second configuration module, configured to align the original segmentation mask and the mirror segmentation mask according to the transformation matrix; A first osteophyte determination module, configured to perform a Boolean operation on the original segmentation mask and the mirror segmentation mask after registration to determine a first osteophyte region; A first osteophyte removal module, configured to remove the first osteophyte region in the original segmentation mask to obtain a candidate segmentation mask; A second osteophyte determination module, configured to determine a second osteophyte region based on the candidate segmentation mask and a preset segmentation mask template; The second osteophyte removal module is used to remove the second osteophyte region in the candidate segmentation mask to obtain a target segmentation mask.

9. An electronic device, characterized in that: The electronic device comprises: processor; a memory for storing instructions executable by the processor; The processor is used to read the executable instructions from the memory and execute the instructions to implement the method described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores instructions, and when the instructions are executed on a terminal device, the terminal device implements the method according to any one of claims 1 to 7.

Images