Segmentation method of lung image, processing device, surgical robot and storage medium

Abstract

The application is suitable for the technical field of lung image processing, and provides a lung image segmentation method, a processing device, a surgical robot and a storage medium, the method comprising: S1 acquiring a lung image; S2 configuring one of first type segmentation methods to segment the lung image to obtain a segmentation result; S3 determining a relative position relationship between a lesion and a reference point based on the lung image and the segmentation result; S4 if the relative position relationship meets a preset condition and the segmentation result does not meet a doctor's diagnosis and treatment requirement, executing step S5; S5 configuring one of segmentation methods which have not been configured from the first type and second type segmentation methods to segment the lung image to obtain a segmentation result; S6 determining an intermediate segmentation result based on at least two of all segmentation results; S7 if the intermediate segmentation result does not meet the doctor's diagnosis and treatment requirement, returning to execute S5, otherwise, taking the intermediate segmentation result as a target segmentation result.

Description

Technical Field

[0001] This application relates to the field of lung and bronchial image processing technology, and more specifically, to a lung image segmentation method, processing device, surgical robot, and storage medium. Background Technology

[0002] Lung biopsy primarily relies on information from CT images and the surgeon's clinical experience. Avoiding the trachea and blood vessels during the procedure significantly reduces the risk of serious complications such as pneumothorax and pulmonary hemorrhage. Using lung images to create a three-dimensional reconstruction of the lungs and trachea provides surgeons with a visual reference for lung biopsy, guiding them in designing precise treatment plans and ensuring effective avoidance of the trachea during the procedure. This significantly reduces complications associated with lung biopsy and has substantial practical value.

[0003] Currently, many methods for lung and bronchial segmentation exist in academia. Among them, lung airway segmentation methods based on region growing are highly sensitive to the selection of thresholds. Taking the dual-threshold region growing method as an example, if the threshold is set too low, it will cause undersegmentation; if it is set too high, it will cause leakage. Deep learning methods, being data-driven, typically require a large amount of training data to achieve excellent results. When the training samples cannot adequately cover the overall population type, the prediction performance is often unsatisfactory. This method has limited segmentation accuracy and requires significant manual intervention. Summary of the Invention

[0004] This application provides a lung image segmentation method, processing device, surgical robot, and storage medium, which can solve the problem of how to better segment lung images.

[0005] In a first aspect, embodiments of this application provide a method for segmenting a lung image, comprising:

[0006] S1: Acquire lung images and determine the type of lung disease based on the lung images;

[0007] S2: If the disease type is the target disease type, configure one of the first type of segmentation methods to segment the lung image and obtain the segmentation result;

[0008] S3: If the segmentation result does not include bronchus with a stratum level not lower than the first preset value, proceed to step S4;

[0009] S4: Select one of the segmentation methods that has not yet been configured from the first and second category segmentation methods to segment the lung image and obtain the segmentation result;

[0010] S5: Determine the intermediate segmentation result based on at least two of all segmentation results;

[0011] S6: If the intermediate segmentation result does not include bronchus with a level not lower than the first preset value, return to execute S4; otherwise, use the intermediate segmentation result as the target segmentation result.

[0012] Secondly, embodiments of this application provide a method for segmenting lung images, including:

[0013] S1: Acquire lung images;

[0014] S2: Configure one of the first-class segmentation methods to segment the lung image and obtain the segmentation result;

[0015] S3: Based on lung images and segmentation results, determine the relative positional relationship between the lesion and the reference point;

[0016] S4: If the relative positional relationship meets the preset conditions, determine whether the relative distance between the bronchus and the lesion in the segmentation result is greater than the first distance threshold, or whether the segmentation result includes bronchus with a layer level not lower than the first preset value. If the relative distance between the bronchus and the lesion in the segmentation result is greater than the first distance threshold, or the segmentation result does not include bronchus with a layer level not lower than the first preset value, proceed to step S5.

[0017] S5: Select one of the segmentation methods that has not yet been configured from the first and second category segmentation methods to segment the lung image and obtain the segmentation result;

[0018] S6: Determine the intermediate segmentation result based on at least two of all segmentation results;

[0019] S7: Determine whether the relative distance between the bronchus and the lesion in the intermediate segmentation result is greater than the first distance threshold, or whether the intermediate segmentation result includes bronchus with a layer level not lower than the first preset value. If the relative distance between the bronchus and the lesion in the intermediate segmentation result is greater than the first distance threshold, or the intermediate segmentation result does not include bronchus with a layer level not lower than the first preset value, return to execute S5; otherwise, use the intermediate segmentation result as the target segmentation result.

[0020] Optionally, the relative positional relationship satisfies preset conditions, including:

[0021] The relative distance between the lesion and the reference point is greater than the second distance threshold;

[0022] or,

[0023] The bronchus at which the lesion is located is at a level greater than a second preset value, and the second preset value is less than a first preset value.

[0024] Optionally, determine the relative positional relationship between the lesion and the reference point, including:

[0025] The location of the lesion was determined based on the lung images;

[0026] The location of the reference point is determined based on the segmentation results;

[0027] Based on the location of the lesion and the location of the reference point, determine the relative distance between the lesion and the reference point or the bronchial level where the lesion is located;

[0028] Optionally, determine the relative positional relationship between the lesion and the reference point, including:

[0029] The location of the lesion was determined based on the lung images;

[0030] The skeleton lines are determined based on the segmentation results;

[0031] Based on the location of the lesion and the skeletal lines, the position of the lesion within the human anatomical structure is determined;

[0032] Based on the location of the lesion in the human anatomical structure, the bronchial level where the lesion is located is determined.

[0033] Optionally, the first type of segmentation method is configured to segment the lung image, and the segmentation result obtained by the second type of segmentation method is configured to segment the lung image, and the segmentation result obtained by the second type of segmentation method is configured to segment the lung image, and the step of determining an intermediate segmentation result based on at least two of all segmentation results includes:

[0034] When the segmentation method configured in the current round is the first type of segmentation method, at least one of the first type of segmentation results and the second type of segmentation results obtained in all previous rounds of segmentation is selected and merged with the first type of segmentation result obtained in the current round of segmentation to obtain the intermediate segmentation result.

[0035] Optionally, the first type of segmentation method is configured to segment the lung image, and the segmentation result obtained by the second type of segmentation method is configured to segment the lung image, and the segmentation result obtained by the second type of segmentation method is configured to segment the lung image, and the step of determining an intermediate segmentation result based on at least two of all segmentation results includes:

[0036] When the segmentation method configured in the current round is the second type of segmentation method, at least one of the first type of segmentation results obtained in all previous rounds of segmentation is selected and merged with the second type of segmentation result obtained in the current round to obtain the intermediate segmentation result.

[0037] Optionally, the first type of segmentation method is configured to segment the lung image, and the segmentation result obtained by the second type of segmentation method is configured to segment the lung image, and the segmentation result obtained by the second type of segmentation method is configured to segment the lung image, and the step of determining an intermediate segmentation result based on at least two of all segmentation results includes:

[0038] When the segmentation method configured in the current round is the second type of segmentation method, at least one of the first type of segmentation results obtained in all previous rounds of segmentation and at least one of the second type of segmentation results obtained in all previous rounds of segmentation are selected and merged with the second type of segmentation result obtained in the current round of segmentation to obtain the intermediate segmentation result.

[0039] Optionally, in step S4, if the relative positional relationship does not meet the preset conditions, the segmentation result is taken as the target segmentation result, and the first type of segmentation method includes the first neural network method.

[0040] Optionally, the first type of segmentation method includes a first neural network method, and the second type of segmentation method includes a second neural network method. The first or second type of neural network method is configured to segment the lung image to obtain a segmentation result, including:

[0041] The lung image is input into the neural network to obtain matrix data, which includes at least one voxel and the confidence level corresponding to the voxel.

[0042] Voxels with confidence levels higher than the first threshold are identified as bronchi, and a target binarized image is obtained as the segmentation result.

[0043] Optionally, if the relative distance between the bronchus and the lesion in the intermediate segmentation result is greater than the first distance threshold, or if the intermediate segmentation result does not include the bronchus with a layer level not lower than the first preset value, and there is no segmentation method that has not yet been configured in the first and second type of segmentation methods, the matrix data obtained by configuring the first or second neural network method to segment the lung image is obtained.

[0044] Lower the first threshold to the second threshold, and identify voxels with confidence levels below the second threshold as bronchi to obtain a new binarized image;

[0045] The new binarized image is XORed with the target binarized image to obtain the XORed binarized image.

[0046] Perform connected component analysis on the binarized graph after XOR to obtain a binarized graph corresponding to multiple connected components;

[0047] Select one candidate connected component from the plurality of connected components;

[0048] The binarized graphs corresponding to the candidate connected components are merged with the target binarized graph to obtain the merged binarized graph;

[0049] When the number of connected components in the merged binarized graph is less than the number of connected components in the target binarized graph, the voxels in the candidate connected components are identified as bronchi, and the candidate connected components are identified as valid connected components.

[0050] The binarized graph of the effective connected region is merged with the target binarized graph to obtain the segmentation result after connecting the broken bronchus.

[0051] Optionally, the first type of segmentation method includes region growing, which is configured to segment the lung image to obtain a segmentation result, including:

[0052] Seed points are selected in the lung images as the starting points for growth;

[0053] Within the neighborhood of the seed point, at least one voxel with an intensity value less than the third threshold is selected;

[0054] The initial value of the dynamic threshold is determined based on the intensity value of the seed point and the intensity value of the selected at least one voxel.

[0055] Region growth is performed based on the seed points, and voxels with initial values ​​less than the dynamic threshold are selected as bronchi to obtain a binarized image as the segmentation result.

[0056] Optionally, region growth is performed based on the initial value of the dynamic threshold, and voxels with intensity values ​​less than the initial value of the dynamic threshold are identified as new seed voxels as bronchi.

[0057] A new threshold is determined based on the initial value of the dynamic threshold and the intensity value of the new seed voxel when the number of new seed voxels reaches a preset number at fixed intervals.

[0058] The new threshold is set as the initial value of the dynamic threshold, and the return step performs region growth based on the initial value of the dynamic threshold.

[0059] Optionally, the lung image is a 3D image, and the step of selecting a seed point in the lung image as the starting point for growth includes:

[0060] A slice is selected from a predetermined location in the lung image, wherein the predetermined location ranges from 0.65 to 0.85.

[0061] Based on the slice threshold, the slice is binarized to obtain a slice binarized image;

[0062] Find the connected components of the sliced ​​binary graph;

[0063] Select a connected region that meets preset conditions, including that the connected region is circular in shape and contains 100 to 500 voxels.

[0064] Select a target voxel from the connected regions that meet the preset conditions, and use the target voxel as a seed point.

[0065] Optionally, the second type of segmentation method includes morphological segmentation, which is configured to segment the lung image to obtain a segmentation result, including:

[0066] The lung image was denoised.

[0067] Based on the intensity values ​​and higher-order thresholds of voxels in the lung image, the denoised lung image is binarized to obtain a binarized image img_a. In the binarized image img_a, voxels with intensity values ​​less than the higher-order threshold include bronchi, while voxels with intensity values ​​greater than the higher-order threshold do not include bronchi.

[0068] Connectivity analysis is performed on the voxel portion of the binarized image img_a that does not include the bronchi, and connected components smaller than the first connected component threshold are removed to obtain the binarized image img_b.

[0069] Perform an opening operation on the binarized image img_b to obtain the binarized image img_c;

[0070] Perform connected component analysis on the binarized graph img_c, and remove connected components that are less than the second connected component threshold to obtain the binarized graph img_d;

[0071] XORing the binarized image img_c and the binarized image img_d yields a higher-level binarized image corresponding to the bronchus.

[0072] Thirdly, embodiments of this application provide a method for segmenting lung images, including:

[0073] S301: Acquire lung images and determine the disease type based on the lung images;

[0074] S302: If the disease is the target disease type, configure one of the first type of segmentation methods to segment the lung image and obtain the segmentation result;

[0075] S303: Based on lung images and segmentation results, determine the relative positional relationship between the lesion and the reference point;

[0076] S304: If the relative positional relationship meets the preset conditions, determine whether the relative distance between the bronchus and the lesion in the segmentation result is greater than the first distance threshold, or whether the segmentation result includes bronchus with a layer level not lower than the first preset value. If the relative distance between the bronchus and the lesion in the segmentation result is greater than the first distance threshold, or the segmentation result does not include bronchus with a layer level not lower than the first preset value, execute step S305.

[0077] S305: Select one of the segmentation methods that has not yet been configured from the first and second category segmentation methods to segment the lung image and obtain the segmentation result;

[0078] S306: Determine an intermediate segmentation result based on at least two of all segmentation results;

[0079] S307: If the relative distance between the bronchus and the lesion in the intermediate segmentation result is greater than the first distance threshold, or if the intermediate segmentation result does not include the bronchus with a layer level not lower than the first preset value, return to execute S305; otherwise, use the intermediate segmentation result as the target segmentation result.

[0080] Fourthly, embodiments of this application provide a method for segmenting lung images, including:

[0081] S401: Acquire lung images;

[0082] S402: Configure one of the first-class segmentation methods to segment the lung image and obtain the segmentation result;

[0083] S403: Based on lung images and segmentation results, determine the relative positional relationship between the lesion and the reference point;

[0084] S404: If the relative positional relationship meets the preset conditions, obtain the lung disease type based on the lung image;

[0085] S405: If the lung disease is the target disease type, determine whether the relative distance between the bronchus and the lesion in the segmentation result is greater than the first distance threshold, or whether the segmentation result includes bronchus with a layer level not lower than the first preset value. If the relative distance between the bronchus and the lesion in the segmentation result is greater than the first distance threshold, or the segmentation result does not include bronchus with a layer level not lower than the first preset value, execute step S406.

[0086] S406: Select one of the segmentation methods that have not yet been configured from the first and second category of segmentation methods to segment the lung image and obtain the segmentation result;

[0087] S407: Determine an intermediate segmentation result based on at least two of all segmentation results;

[0088] S408: If the relative distance between the bronchus and the lesion in the intermediate segmentation result is greater than the first distance threshold, or if the intermediate segmentation result does not include the bronchus with a layer level not lower than the second preset value, return to execute S406; otherwise, use the intermediate segmentation result as the target segmentation result.

[0089] Fifthly, embodiments of this application provide a method for segmenting lung images, including:

[0090] S501: Acquire lung images and determine the type of lung disease based on the lung images;

[0091] S502: Configure one of the first-class segmentation methods to segment the lung image and obtain the segmentation result;

[0092] S503: Based on lung images and segmentation results, determine the relative positional relationship between the lesion and the reference point.

[0093] S504: If the disease type is the target disease type and the relative positional relationship meets the preset conditions, then further determine whether the relative distance between the bronchus and the lesion in the segmentation result is greater than the first distance threshold, or whether the segmentation result includes bronchus with a layer level not lower than the first preset value. If the relative distance between the bronchus and the lesion in the segmentation result is greater than the first distance threshold, or the segmentation result does not include bronchus with a layer level not lower than the first preset value, then execute step S505.

[0094] S505: Select one of the segmentation methods that has not yet been configured from the first and second category segmentation methods to segment the lung image and obtain the segmentation result;

[0095] S506: Determine an intermediate segmentation result based on at least two of all segmentation results;

[0096] S507: If the relative distance between the bronchus and the lesion in the intermediate segmentation result is greater than the first distance threshold, or if the intermediate segmentation result does not include the bronchus with a layer level not lower than the first preset value, return to execute S505; otherwise, use the intermediate segmentation result as the target segmentation result.

[0097] Sixthly, embodiments of this application provide an image processing apparatus, including:

[0098] A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the lung image segmentation method of any one of the first to fifth aspects described above when executing the computer program.

[0099] In a seventh aspect, embodiments of this application provide a surgical robot, including: a main operating table, a slave operating device; and an image processing device coupled to the main operating table and the slave operating device, and configured to perform the lung image segmentation method of any one of the first to fifth aspects described above.

[0100] Eighthly, embodiments of this application provide a computer-readable storage medium storing a computer program, which is executed by a processor as the lung image segmentation method of any one of the first to fifth aspects described above.

[0101] Ninthly, embodiments of this application provide a computer program product that, when run on a terminal device, causes the terminal device to execute the lung image segmentation method of any one of the first to fifth aspects described above.

[0102] It is understood that the beneficial effects of aspects two through nine above can be found in the relevant descriptions in aspect one above, and will not be repeated here.

[0103] The beneficial effects of this application embodiment compared with the prior art are as follows: This application embodiment acquires a lung image through S1; S2 configures one of the first type of segmentation methods to segment the lung image and obtains a segmentation result; S3 determines the relative positional relationship between the lesion and the reference point based on the lung image and the segmentation result; S4 if the relative positional relationship meets the preset conditions, and it is determined that the segmentation result does not meet the doctor's diagnostic and treatment requirements, then execute step S5; S5 configures one of the segmentation methods that have not yet been configured from the first and second types of segmentation methods to segment the lung image and obtains a segmentation result; S6 determines an intermediate segmentation result based on at least two of all segmentation results; S7 if it is determined that the intermediate segmentation result does not meet the doctor's diagnostic and treatment requirements, then return to execute S5; otherwise, the intermediate segmentation result is used as the target segmentation result. By adopting appropriate segmentation methods for different lesion conditions, the automation level, segmentation efficiency, and segmentation accuracy of segmentation are improved, thus meeting the doctor's diagnostic and treatment needs. Attached Figure Description

[0104] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0105] Figure 1 This is a schematic diagram of a lung image segmentation method according to an embodiment of this application;

[0106] Figure 2This is a schematic diagram of another lung image segmentation method in an embodiment of this application;

[0107] Figures 3a-3e This is a schematic diagram illustrating a lung image segmentation method according to an embodiment of this application;

[0108] Figures 4a to 4d This is a schematic diagram illustrating another lung image segmentation method in an embodiment of this application;

[0109] Figures 5a-5c This is a schematic diagram of the mask extraction process in an embodiment of this application;

[0110] Figure 6 This is a schematic diagram of the image processing device in an embodiment of this application. Detailed Implementation

[0111] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present application will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0112] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0113] Figure 1 This is a schematic diagram of an embodiment of the lung image segmentation method provided by the present invention, the method comprising:

[0114] Step S1: Acquire lung images and determine the type of lung disease based on the lung images;

[0115] If the disease type is the target disease type, determine to perform multi-round combination segmentation;

[0116] If the disease type is not the target disease type, a single-round segmentation will be performed.

[0117] The target disease types include those that affect imaging quality and cause extensive inflammation, such as emphysema, bronchiectasis, and pneumonia; the non-target disease types include those that do not affect imaging quality and do not cause extensive inflammation, such as pulmonary nodules.

[0118] Among these methods, obtaining the disease type of the lungs can optionally include identifying the disease type of the organ based on the morphological characteristics and classification algorithms of the lungs in the prior art, which is not limited in this embodiment; or directly obtaining the disease type information input by the user, such as the disease type information input by the user in the interactive interface.

[0119] Because animals and humans differ significantly, differentiation is currently necessary. Optionally, features in the lung image can be extracted from the lung mask to determine whether it's an animal or a human, or information input by the user—such as the user entering animal or human information on the interface—can be used. If it is an animal, a combined segmentation approach is employed. Understandably, if sufficient samples of animal lung images are collected and effectively trained to form a neural network specific to animals, then for non-target disease types in animals, a single-round segmentation using this animal-specific neural network can be used, while for target disease types in animals, a multi-round combined segmentation approach can be employed.

[0120] Step S2: If the disease type is the target disease type, a multi-round combined segmentation method is used. First, one of the first-class segmentation methods is configured to segment the lung image to obtain the segmentation result. The first-class segmentation method is used to cut the whole bronchus, which can also be called the main bronchus. Optionally, one of the first-class segmentation methods can be configured to segment the lung image first to obtain the segmentation result, and then it can be determined whether the disease type is the target disease type. If it is the target disease type, proceed to step S3.

[0121] Step S3: Determine whether the segmentation result meets the doctor's diagnostic and treatment requirements. For example, if it includes bronchus with a layer level not lower than the first preset value, the segmentation result is taken as the target segmentation result; if the segmentation result does not include bronchus with a layer level not lower than the first preset value, proceed to step S4.

[0122] The process determines whether the segmentation result includes bronchi at a level no lower than the first preset value, i.e., whether the segmentation result meets the doctor's usage requirements. If the segmentation result includes bronchi at a level no lower than the first preset value, the segmentation result meets the doctor's usage requirements; if the segmentation result does not include bronchi at a level no lower than the first preset value, the segmentation result does not meet the doctor's usage requirements. The first preset value can be flexibly set by the doctor, such as level 5 or level 6. If the segmentation result does not meet the usage requirements, step S4 needs to be executed to continue the next round of segmentation.

[0123] Optionally, determining whether the segmentation result meets the doctor's diagnostic and treatment requirements may also include determining whether the relative distance between the bronchus and the lesion in the segmentation result is greater than a first distance threshold, specifically including:

[0124] Determining the location of lesions based on lung images;

[0125] Determine the skeleton lines based on the segmentation results;

[0126] Based on the location of the lesion and the skeletal line, the relative distance between the bronchus and the lesion in the segmentation result is determined. For example, the skeletal line includes multiple points, and the minimum distance between the lesion and the points on the skeletal line is used as the relative distance between the bronchus and the lesion in the segmentation result.

[0127] The system determines whether the relative distance between the bronchus and the lesion in the segmentation result is greater than a first distance threshold. If it is greater than the first distance threshold, the segmentation result does not meet the doctor's diagnostic and treatment requirements; if it is not greater than the first distance threshold, the segmentation result meets the doctor's diagnostic and treatment requirements.

[0128] Optionally, the doctor's diagnostic and treatment requirements can be determined based on two conditions: whether the segmentation result includes a bronchus with a level not lower than the first preset value and whether the relative distance between the bronchus and the lesion in the segmentation result is greater than the first distance threshold. Only when both conditions are met is the doctor's diagnostic and treatment requirements considered to be met.

[0129] Step S4: Select one of the segmentation methods that has not yet been configured from the first and second type of segmentation methods to segment the lung image and obtain the segmentation result. The second type of segmentation method is used to cut the higher-level bronchi.

[0130] Optionally, the same segmentation method may not be used repeatedly to segment the lung image. Therefore, one of the segmentation methods that has not yet been configured needs to be selected to segment the lung image to obtain the segmentation result. Either the first type of segmentation method or the second type of segmentation method can be selected.

[0131] Step S5: Determine the intermediate segmentation result based on at least two of all segmentation results.

[0132] When multiple rounds of segmentation are performed, the segmentation results of at least two rounds can be selected and merged to determine the intermediate segmentation result.

[0133] Step S6: Determine whether the intermediate segmentation result meets the doctor's diagnostic and treatment requirements. For example, if it does not include bronchus with a layer level not lower than the first preset value, return to execute S4; otherwise, use the intermediate segmentation result as the target segmentation result.

[0134] Optionally, determining whether the intermediate segmentation results meet the doctor's diagnostic and treatment requirements may also include:

[0135] The system determines whether the relative distance between the bronchus and the lesion in the intermediate segmentation result is greater than a first distance threshold. If it is greater than the first distance threshold, the intermediate segmentation result does not meet the doctor's diagnostic and treatment requirements; if it is not greater than the first distance threshold, the intermediate segmentation result meets the doctor's diagnostic and treatment requirements. The implementation principle is the same as the aforementioned determination of whether the relative distance between the bronchus and the lesion in the segmentation result is greater than the first distance threshold, and will not be repeated here.

[0136] Optionally, the doctor's diagnostic and treatment requirements can be determined based on two conditions: whether the intermediate segmentation result includes a bronchus with a level not lower than the first preset value and whether the relative distance between the bronchus and the lesion in the intermediate segmentation result is greater than the first distance threshold. Only when both conditions are met is the doctor's diagnostic and treatment requirements considered to be met.

[0137] When multiple merged intermediate segmentation results all meet the doctor's diagnostic and treatment requirements, such as including bronchi at a level not lower than a first preset value, one can be selected for the doctor's use. The selection criteria can be flexibly set; for example, the higher the level of bronchi included, the better; or the more complete the bronchi at a particular level the doctor is interested in, the better; the longer the skeleton line, the better; the more branches of the skeleton line, the better; and the closer the relative distance between the bronchi and the lesion, the better. Alternatively, the doctor can observe the merged results and select one as the target segmentation result, or randomly select one as the target segmentation result. The length of the skeleton line, the number of branches of the skeleton line, and the relative distance between the bronchi and the lesion can be determined using methods in the prior art based on the intermediate segmentation results; this application does not impose any limitations on these aspects.

[0138] This application embodiment segments lung images according to the disease type of the lung image, thereby achieving adaptive use of appropriate segmentation methods for different lung images. This allows for the acquisition of lung and bronchial segments, improving the automation, efficiency, and accuracy of segmentation, and meeting the diagnostic and treatment needs of doctors.

[0139] Furthermore, if the disease type is a non-target disease type, a single-round segmentation is performed. This involves configuring one of the first-class segmentation methods to segment the lung image, obtaining the target segmentation result. The first-class segmentation method includes a first neural network method. For example, the lung image is input into a neural network to obtain matrix data, which includes at least one voxel and its corresponding confidence score. Voxels with confidence scores higher than a first preset threshold are identified as bronchioles to obtain a target binary image. This target binary image is then used as the segmentation result. The confidence score represents the probability that each voxel in the lung image is a bronchiole. For example, the first preset threshold is 0.3. Voxels with confidence scores lower than 0.3 are not bronchioles, while those with confidence scores higher than 0.3 are identified as bronchioles. Voxels identified as bronchioles can be assigned a value of 1, and those not identified as bronchioles can be assigned a value of 0. Alternatively, voxels identified as bronchioles can also be assigned a value of 0, and those not identified as bronchioles can be assigned a value of 1, thus obtaining the target binary image. Figure 2 The image shown is a segmentation result obtained through the first neural network method. The first neural network includes 3DUnet, AirwayNet, and BronchusNet. This first neural network is trained using multiple sample data and is suitable for segmenting the entire bronchus. Increasing the diversity of samples during training improves the adaptability of the first neural network. The specific training process can employ existing technologies, and this application does not impose specific limitations on it.

[0140] Furthermore, the appropriate segmentation method can be selected based on the order recommended by big data statistical results, instead of configuring one of the segmentation methods that has not yet been configured in the first and second categories of segmentation methods as in step S4, thereby improving the segmentation effect; or the combination order of segmentation methods can be determined according to the doctor's instructions, for example, the doctor can input the combination order of segmentation methods through the human-computer interface.

[0141] This application embodiment determines whether to perform single-round segmentation or combined segmentation on lung images based on the disease characteristics of the lung images. This enables the adaptive use of appropriate segmentation methods for different lung images, improving the automation, efficiency, and accuracy of segmentation, and meeting the diagnostic and treatment needs of doctors.

[0142] In another embodiment of the lung image segmentation method provided in this application, such as... Figure 2 This is a schematic diagram of another embodiment of the lung image segmentation method provided by the present invention. The lung image is segmented using a suitable segmentation method based on the lesions, specifically including:

[0143] S201: Acquire lung images.

[0144] S202: Configure one of the first-class segmentation methods to segment the lung image and obtain the segmentation result.

[0145] S203: Based on lung images and segmentation results, determine the relative positional relationship between the lesion and the reference point.

[0146] Determine the relative position of the lesion and the reference point, including: determining the relative distance between the lesion and the reference point or the bronchial level at which the lesion is located.

[0147] Determining the relative distance between the lesion and the reference point includes:

[0148] The location of the lesion is determined based on lung images; the location of a reference point is determined based on the segmentation results; optionally, the location of the lesion is determined based on lung images; the location of a reference point is determined based on the segmentation results, for example, the main carina can be selected as the reference point. The location of the lesion can be determined based on lung images, and existing technologies can be used specifically; this application does not impose any limitations. The location of the reference point can be determined based on segmentation results, and existing technologies can be used specifically; this application does not impose any limitations.

[0149] The relative distance between the lesion and the reference point is determined based on their locations.

[0150] Determining the bronchial level where the lesion is located includes:

[0151] The relative distance between the lesion and the reference point can be used to determine the bronchial level where the lesion is located. For example, based on statistics of a large number of samples, there is a certain correspondence between the relative distance value and the bronchial level where the lesion is located.

[0152] Alternatively, determine the bronchial level at which the lesion is located, including:

[0153] Determining the location of lesions based on lung images;

[0154] Determine the skeleton lines based on the segmentation results;

[0155] Based on the location of the lesion and the skeletal line, the position of the lesion on the human anatomical structure can be determined. For example, the relative position of the point on the skeletal line closest to the lesion can be used to determine the position of the lesion on the human anatomical structure.

[0156] Based on the location of the lesion in the human anatomy, such as the upper or middle lobe of the lung, the bronchial level where the lesion is located is determined.

[0157] S204: If the relative positional relationship meets the preset conditions, determine whether the relative distance between the bronchus and the lesion in the segmentation result is greater than the first distance threshold, or whether the segmentation result includes bronchus with a layer level not lower than the first preset value. If the relative distance between the bronchus and the lesion in the segmentation result is greater than the first distance threshold, or the segmentation result does not include bronchus with a layer level not lower than the first preset value, execute step S205.

[0158] Among them, the relative positional relationship meets the preset conditions, including:

[0159] The relative distance between the lesion and the reference point is greater than the second distance threshold;

[0160] or,

[0161] The bronchial level where the lesion is located is greater than the second preset value, and the second preset value is less than the first preset value.

[0162] If the relative distance between the lesion and the reference point is greater than the second distance threshold, the lesion can be determined to be near a more distant bronchus, and therefore segmentation of the higher-level bronchus is sufficient. If the relative distance is less than the second distance threshold, the lesion can be determined to be near a more recent bronchus, and segmentation of the lower-level bronchus is sufficient.

[0163] When the relative distance between the bronchus and the lesion in the segmentation result is greater than the first distance threshold, it indicates that the segmentation effect of the bronchi around the lesion is not ideal; when the relative distance between the bronchus and the lesion is not greater than the first distance threshold, it indicates that the segmentation effect of the bronchi around the lesion is ideal. When the segmentation result does not include bronchi at or above the first preset value, it is considered that the segmentation effect of the higher-level bronchus is not ideal; when the segmentation result includes bronchi at or above the first preset value, it is considered that the segmentation effect of the higher-level bronchus is ideal. When the segmentation effect is not ideal, it is necessary to return to S205 for the next round of segmentation. When the segmentation effect is ideal, the segmentation result is taken as the target segmentation result, and the segmentation ends.

[0164] Optionally, determining whether the segmentation result meets the doctor's diagnostic and treatment requirements may also include determining whether the relative distance between the bronchus and the lesion in the segmentation result is greater than a first distance threshold, including:

[0165] Determining the location of lesions based on lung images;

[0166] Determine the skeleton lines based on the segmentation results;

[0167] Based on the location of the lesion and the skeletal line, the relative distance between the bronchus and the lesion in the segmentation result is determined. For example, the skeletal line includes multiple points, and the minimum distance between the lesion and the points on the skeletal line is used as the relative distance between the bronchus and the lesion in the segmentation result.

[0168] The system determines whether the relative distance between the bronchus and the lesion in the segmentation result is greater than a first distance threshold. If it is greater than the first distance threshold, the segmentation result does not meet the doctor's diagnostic and treatment requirements; if it is not greater than the first distance threshold, the segmentation result meets the doctor's diagnostic and treatment requirements.

[0169] Optionally, the doctor's diagnostic and treatment requirements can be determined based on two conditions: whether the segmentation result includes a bronchus with a level not lower than the first preset value and whether the relative distance between the bronchus and the lesion in the segmentation result is greater than the first distance threshold. Only when both conditions are met is the doctor's diagnostic and treatment requirements considered to be met.

[0170] S205: Select one of the segmentation methods that has not yet been configured from the first and second category segmentation methods to segment the lung image and obtain the segmentation result;

[0171] S206: Determine the intermediate segmentation result based on at least two of all segmentation results.

[0172] S207: Determine whether the intermediate segmentation result meets the doctor's diagnostic and treatment requirements. For example, whether the relative distance between the bronchus and the lesion in the intermediate segmentation result is greater than the first distance threshold, or whether the intermediate segmentation result includes bronchus with a layer level not lower than the first preset value. If the relative distance between the bronchus and the lesion in the intermediate segmentation result is greater than the first distance threshold, or the intermediate segmentation result does not include bronchus with a layer level not lower than the first preset value, i.e., it does not meet the doctor's requirements, return to execute S205; otherwise, use the intermediate segmentation result as the target segmentation result.

[0173] When the relative distance between the bronchus and the lesion in the intermediate segmentation result is greater than the first distance threshold, it indicates that the segmentation effect of the bronchus around the lesion is not ideal; when the relative distance between the bronchus and the lesion is not greater than the first distance threshold, it indicates that the segmentation effect of the bronchus around the lesion is ideal. When the intermediate segmentation result does not include bronchus segments not lower than the first preset value, it is considered that the segmentation effect of the higher-level bronchus is not ideal; when the intermediate segmentation result includes bronchus segments not lower than the first preset value, it is considered that the segmentation effect of the higher-level bronchus is ideal. When the segmentation effect is not ideal, it is necessary to return to S205 for the next round of segmentation. When the segmentation effect is ideal, the intermediate segmentation result is used as the target segmentation result, and the segmentation ends. The principle of determining whether the relative distance between the bronchus and the lesion in the intermediate segmentation result is greater than the first distance threshold is the same as the aforementioned principle of determining whether the relative distance between the bronchus and the lesion in the segmentation result is greater than the first distance threshold, and will not be repeated here.

[0174] Optionally, the doctor's diagnostic and treatment requirements can be determined based on two conditions: whether the intermediate segmentation result includes a bronchus with a level not lower than the first preset value and whether the relative distance between the bronchus and the lesion in the segmentation result is greater than the first distance threshold. Only when both conditions are met is the doctor's diagnostic and treatment requirements considered to be met.

[0175] This application embodiment selects an appropriate segmentation method based on the lesion to segment the lung image, thereby achieving adaptive use of the appropriate segmentation method for different lung images, improving the automation, efficiency and accuracy of segmentation, and obtaining the lung bronchi to meet the diagnostic and treatment needs of doctors.

[0176] In another embodiment of the lung image segmentation method provided in this application, the disease type is first determined based on the lung image, and then the need for further acquisition of lesions is determined according to the situation, specifically including:

[0177] S301: Acquire lung images and determine the disease type based on the lung images;

[0178] S302: If the disease is the target disease type, configure one of the first type of segmentation methods to segment the lung image and obtain the segmentation result;

[0179] S303: Based on lung images and segmentation results, determine the relative positional relationship between the lesion and the reference point. The implementation principle is as described above and will not be repeated here.

[0180] S304: If the relative positional relationship meets the preset conditions, determine whether the segmentation result meets the doctor's diagnostic and treatment requirements. For example, whether the relative distance between the bronchus and the lesion in the segmentation result is greater than the first distance threshold, or whether the segmentation result includes bronchus with a layer level not lower than the first preset value. If the doctor's diagnostic and treatment requirements are not met, proceed to step S305.

[0181] S305: Select one of the segmentation methods that has not yet been configured from the first and second category segmentation methods to segment the lung image and obtain the segmentation result;

[0182] S306: Determine an intermediate segmentation result based on at least two of all segmentation results;

[0183] S307: Determine whether the intermediate segmentation result meets the doctor's diagnostic and treatment requirements. For example, whether the relative distance between the bronchus and the lesion in the intermediate segmentation result is greater than the first distance threshold, or whether the intermediate segmentation result includes bronchus with a layer level not lower than the second preset value. If it does not meet the doctor's diagnostic and treatment requirements, return to execute S305; otherwise, use the intermediate segmentation result as the target segmentation result.

[0184] Furthermore, if in step S304 the relative positional relationship does not meet the preset conditions, the segmentation result is taken as the target segmentation result. The first type of segmentation method includes the first neural network method. That is, when the disease type is the target disease type, multi-round combination segmentation is not directly used, but the situation of the lesion is further judged. For example, if the relative positional relationship between the lesion and the reference point meets the preset conditions, multi-round combination segmentation is performed, thereby improving the segmentation efficiency.

[0185] This application embodiment selects an appropriate segmentation method for lung images based on lesions and disease types, thereby achieving adaptive use of appropriate segmentation methods for different lung images, improving the automation, efficiency, and accuracy of segmentation, and meeting the diagnostic and treatment needs of doctors.

[0186] In another embodiment of the lung image segmentation method provided in this application, the lesion condition is first obtained, and then the segmentation method is determined based on whether the disease type needs to be obtained. Specifically, it includes:

[0187] S401: Acquire lung images;

[0188] S402: Configure one of the first-class segmentation methods to segment the lung image and obtain the segmentation result;

[0189] S403: Based on lung images and segmentation results, determine the relative positional relationship between the lesion and the reference point. The implementation principle is as described above and will not be repeated here.

[0190] S404: If the relative positional relationship meets the preset conditions, obtain the lung disease type based on the lung image;

[0191] S405: If the lung disease is the target disease type, further determine whether the segmentation result meets the doctor's diagnostic and treatment requirements. For example, whether the relative distance between the bronchus and the lesion in the segmentation result is greater than the first distance threshold, or whether the segmentation result includes bronchus with a layer level not lower than the first preset value. If it does not meet the doctor's diagnostic and treatment requirements, proceed to step S406.

[0192] In this embodiment of the application, when the relative positional relationship between the lesion and the reference point meets the preset conditions, instead of directly using multi-round combination segmentation, the disease type is further determined. Multi-round combination segmentation is only performed when it is the target disease type, and single-round segmentation is performed when it is not the target disease type, thereby improving the segmentation efficiency.

[0193] S406: Select one of the segmentation methods that have not yet been configured from the first and second category of segmentation methods to segment the lung image and obtain the segmentation result;

[0194] S407: Determine an intermediate segmentation result based on at least two of all segmentation results;

[0195] S408: Determine whether the intermediate segmentation result meets the doctor's diagnostic and treatment requirements. For example, determine whether the relative distance between the bronchus and the lesion in the intermediate segmentation result is greater than the first distance threshold, or whether the intermediate segmentation result includes bronchus with a layer level not lower than the second preset value. If the results do not meet the doctor's diagnostic and treatment requirements, return to execute S406; otherwise, use the intermediate segmentation result as the target segmentation result.

[0196] This application embodiment selects an appropriate segmentation method for lung images based on lesions and disease types, thereby achieving adaptive use of appropriate segmentation methods for different lung images, improving the automation, efficiency, and accuracy of segmentation, and meeting the diagnostic and treatment needs of doctors.

[0197] In another embodiment of the lung image segmentation method provided in this application, the segmentation method is determined based on the condition of the lesions and the type of disease, including:

[0198] S501: Acquire lung images and determine the type of lung disease based on the lung images;

[0199] S502: Configure one of the first-class segmentation methods to segment the lung image and obtain the segmentation result;

[0200] S503: Based on the lung images and segmentation results, determine the relative positional relationship between the lesion and the reference point. It is understood that determining the type of lung disease based on the lung images may be done in step 503, not in step S501.

[0201] S504: If the disease type is the target disease type and the relative positional relationship meets the preset conditions, then further determine whether the segmentation result meets the doctor's diagnostic and treatment requirements. For example, whether the relative distance between the bronchus and the lesion in the segmentation result is greater than the first distance threshold, or whether the segmentation result includes bronchus with a layer level not lower than the first preset value. If it does not meet the doctor's diagnostic and treatment requirements, proceed to step S505.

[0202] S505: Select one of the segmentation methods that has not yet been configured from the first and second category segmentation methods to segment the lung image and obtain the segmentation result;

[0203] S506: Determine an intermediate segmentation result based on at least two of all segmentation results;

[0204] S507: Determine whether the intermediate segmentation result meets the doctor's diagnostic and treatment requirements. For example, whether the relative distance between the bronchus and the lesion in the intermediate segmentation result is greater than the first distance threshold, or whether the intermediate segmentation result includes bronchus with a layer level not lower than the first preset value. If it does not meet the doctor's diagnostic and treatment requirements, return to execute S505; if it meets the doctor's diagnostic and treatment requirements, use the intermediate segmentation result as the target segmentation result.

[0205] This application embodiment selects an appropriate segmentation method for lung images based on lesions and disease types, thereby achieving adaptive use of appropriate segmentation methods for different lung images, improving the automation, efficiency, and accuracy of segmentation, and meeting the diagnostic and treatment needs of doctors.

[0206] In another embodiment of the lung image segmentation method provided in this application, the first type of segmentation method includes a region growing method, which is configured to segment the lung image to obtain a segmentation result, including:

[0207] Seed points are selected in the lung image. Based on a dynamic threshold, region growing is performed using the seed points. Voxels with intensity values ​​less than the dynamic threshold are identified as bronchi, resulting in a binarized image as the segmentation result. Specifically, this includes:

[0208] Step 1: Select a seed point as the starting point for growth. There are three methods for selecting the seed point:

[0209] Method 1: Select a slice at a preset location in the lung image. The lung image is a 3D image and includes multiple slices. Optionally, one slice can be selected for binarization. The slice position can optionally be within the range of 0.65 to 0.85. For example, with the top of the lung as position 0 and the bottom of the lung as position 1, the slice corresponding to position 0.8 is selected as the analysis object.

[0210] Optionally, a slice threshold is preset. Based on the slice threshold, the slices are binarized to obtain a slice binarized image. For example, if the intensity value of a voxel in a slice is less than the slice threshold, the voxel is identified as a bronchus. Each voxel in the image can use an intensity value to distinguish different types of tissues, fluids, structures, etc., within the image space. For example, when the image data is CT image data, each voxel is associated with a Hounsfield value, abbreviated as HU value. Other intensity values ​​can be used in other embodiments.

[0211] Find the connected components of the sliced ​​binary graph;

[0212] Connected components that meet preset conditions are selected. These conditions include a circular shape and a number of voxels less than 500 and greater than 100 (i.e., 500 > number of voxels > 100). One voxel from the connected component is then selected as the seed point. The circle does not need to be perfectly circular; any voxel close to a circle can be considered circular. In this embodiment, the voxel selected as the seed point is called the target voxel, and any voxel from the selectable connected component can be used as the target voxel.

[0213] The second approach: If the region growing method is not configured in the first round of segmentation, then one of the voxels identified as a bronchus from the segmentation results obtained in all previous rounds of segmentation can be selected as the seed point.

[0214] The third method: Manually select seed points. Users can observe whether there are areas with circular, dark holes that appear consecutively in adjacent slices, and select such areas as seed points through input devices, such as interactive interfaces.

[0215] Step 2: Within the neighborhood of the selected seed point, select at least one voxel that satisfies the preset constraints.

[0216] Preset constraints include two aspects:

[0217] The distance from the seed point should not be too far; voxels within a 5*5*5 neighborhood can be selected.

[0218] The HU intensity value of the voxel is less than the empirical threshold.

[0219] Step 3: Determine the initial value of the dynamic threshold based on the intensity value of the selected seed point and the intensity value of at least one selected voxel.

[0220] Optionally, the average of the intensity value of the seed point and the intensity values ​​of at least one selected voxel can be calculated, and this average value can be used as the initial value of the dynamic threshold. In this embodiment, the intensity value of the seed point is denoted as HU. 种子， The intensity value of at least one selected voxel is denoted as HU. i , i represents the index of at least one selected voxel, the number of selected voxels is m, and i is a natural number from 1 to m.

[0221] Furthermore, the intensity value of the seed point is assigned a weight α, the average intensity value of the selected voxels is assigned a weight β, and then the weighted intensity values ​​are averaged to obtain the initial value of the dynamic threshold.

[0222] Step 4: Update the initial value of the dynamic threshold.

[0223] Region growth is performed based on the initial value of the dynamic threshold. Voxels with intensity values ​​less than the initial value are identified as new seed voxels and then identified as bronchi.

[0224] Based on the initial value of the dynamic threshold and the intensity value of the new seed voxel, the initial value is updated when preset conditions are met, such as at fixed intervals or when the number of new seed voxels reaches a preset number.

[0225] Set the updated initial value as the initial value for the dynamic threshold, and return to the step to perform region growth based on the initial value.

[0226] In this embodiment, voxels with intensity values ​​less than an initial threshold are searched, and the initial threshold is updated when a preset condition is met. The search continues with the updated initial threshold to find voxels with intensity values ​​less than the updated initial threshold, and the threshold is updated again when the preset condition is met. This process is repeated, continuously updating the threshold and searching for voxels with intensity values ​​less than the new threshold. This allows for dynamic adjustment of the threshold based on newly added voxels, thereby utilizing the global information of voxels on the lung airway tree to dynamically adjust the threshold and overcoming the shortcomings of using fixed thresholds for region growth in the prior art.

[0227] In another embodiment of the lung image segmentation method provided in this application, it further includes: connecting the broken bronchus, specifically including:

[0228] If the relative distance between the bronchus and the lesion in the intermediate segmentation result is greater than the first distance threshold, or if the intermediate segmentation result does not include bronchus at a level not lower than the first preset value, and there is no unconfigured segmentation method in either the first or second type of segmentation method, the broken bronchus can be connected. Connecting broken bronchus is applicable to segmentation results obtained by the first or second neural network method in all previous rounds of segmentation. Optionally, other conditions for connecting broken bronchus can also be set, such as the number of segmentation rounds reaching a preset number, or the segmentation time reaching a preset time. Like the first neural network, the second neural network is also obtained after training using multiple sample data, and is suitable for cutting high-level bronchus. The network parameters of the first neural network and the second neural network are different.

[0229] Optionally, when the segmentation method configured for all previous rounds is the first or second neural network method, it includes: inputting the lung image into the neural network to obtain matrix data, wherein the matrix data is three-dimensional volume data, including at least one voxel and the confidence level corresponding to the voxel;

[0230] Voxels with confidence scores higher than the first threshold are identified as bronchi, and the resulting binarized target image is used as the segmentation result.

[0231] Furthermore, connecting the broken bronchus in the target binarized image includes:

[0232] Step 1: Reduce the first threshold in the segmentation method configured for all previous rounds (either the first or second neural network method) to the second threshold, and re-obtain the binarized image. For example, if the first neural network method is selected, reduce the threshold set in the first neural network method to the second threshold; if the second neural network method is selected, reduce the threshold set in the second neural network method to the second threshold.

[0233] Step 2: Perform an XOR operation between the newly obtained binarized image and the target binarized image corresponding to the segmentation result obtained from one of the selected first or second neural network methods to obtain the XORed binarized image, thereby obtaining the bronchial portion obtained due to the reduction of the threshold.

[0234] Step 3: Perform connected component analysis on the binarized graph after XOR to obtain a binarized graph corresponding to multiple connected components;

[0235] Step 4: Select one candidate connected component from multiple connected components. Smaller connected components can be ignored. Therefore, only all or part of the connected components with a volume greater than the threshold can be selected as candidate connected components.

[0236] The binarized graphs corresponding to the candidate connected components are merged with the target binarized graphs corresponding to the segmentation results obtained from one of the selected first or second neural network methods to obtain the merged binarized graphs.

[0237] When the number of connected components in the merged binarized graph is less than the number of connected components in the target binarized graph corresponding to the segmentation result obtained by one of the selected first or second neural network methods, the voxels in the corresponding candidate connected components are identified as bronchi to connect the broken bronchi.

[0238] Repeat step four to determine whether each voxel in the candidate connected component is identified as a bronchus. If so, the corresponding connected component is a valid connected component.

[0239] Furthermore, the binarized graphs of all or part of the effective connected components are merged with the target binarized graph corresponding to the segmentation result obtained from one of the selected first or second neural network methods to obtain the segmentation result after connecting the broken bronchus; or the binarized graph obtained by extracting the largest connected component from the merged segmentation result is used as the segmentation result after connecting the broken bronchus.

[0240] The segmentation results obtained by connecting the fractured bronchus are updated using the segmentation results of all previous rounds. The segmentation method configured is either the first or the second neural network method. Then, the intermediate segmentation results and the target segmentation results are determined based on the updated segmentation results.

[0241] If lowering the threshold of either the first or second neural network method is insufficient, the threshold of the other first or second neural network method can be lowered to optimize the segmentation results in each round and the intermediate segmentation results after merging.

[0242] Optionally, after the segmentation is completed using the first or second neural network method, connected component analysis can be performed. If there is a break between the largest connected component and a connected component exceeding a certain threshold, it is determined that a bronchus is broken, and the aforementioned steps to connect the broken bronchus can be executed. The principle is the same and will not be elaborated further. The threshold can be flexibly set based on experience.

[0243] Optionally, the doctor may be instructed to connect the severed bronchus, and the aforementioned steps for connecting the severed bronchus may be performed.

[0244] This application embodiment improves the automation level and segmentation effect of lung images by selecting a first threshold and setting a second threshold to lower the configured segmentation method to either a first or a second neural network method, and automatically connecting broken bronchi.

[0245] In another embodiment provided in this application, the configurable segmentation method includes a first type of segmentation method and a second type of segmentation method. The segmentation result obtained using the first type of segmentation method is the first type of segmentation result, and the segmentation result obtained using the second type of segmentation method is the second type of segmentation result. Optionally, an intermediate segmentation result is determined based on at least two of all segmentation results, including:

[0246] When the segmentation method configured for the current round is the first type segmentation method, at least one of the first type segmentation results and the second type segmentation results obtained from all previous rounds of segmentation is selected and merged with the first type segmentation result obtained in the current round to obtain the intermediate segmentation result of the current round.

[0247] When the segmentation method configured for the current round is the second type of segmentation method, at least one of the first type of segmentation results obtained from all previous rounds of segmentation is selected and merged with the second type of segmentation result obtained in the current round to obtain the intermediate segmentation result of the current round.

[0248] When the segmentation method configured for the current round is the second type of segmentation method, at least one of the first type of segmentation results obtained from all previous rounds of segmentation and at least one of the second type of segmentation results obtained from all previous rounds of segmentation are selected and merged with the second type of segmentation result obtained in the current round to obtain the intermediate segmentation result of the current round.

[0249] When the lung disease type is the target disease type, using only one segmentation method for a single round of segmentation on the lung image is insufficient to meet the diagnostic and treatment needs of doctors. Therefore, multiple segmentation methods are used for multiple rounds of segmentation, and then the results of each round are merged. Since different segmentation methods are used in multiple rounds, such as some using the first type of segmentation method and others using the second type, the merging methods for the results of each round also vary. For example, the results of the previous round (round 1, 2, 3, or round i) may be selected and merged with the results of the current round. It should be noted that at least one round using the first type of segmentation method must be selected. In this embodiment, the total number of segmentation rounds is denoted as i+1. The specific number of rounds selected for merging with the results of the current round includes several cases:

[0250] Starting with the fewest rounds, if the merged intermediate segmentation results meet the doctor's diagnostic and treatment requirements, they are used as the target segmentation result; if none of the merged intermediate segmentation results meet the doctor's diagnostic and treatment requirements, the number of rounds is increased. For example, the segmentation results of round 1 and the current round are selected for merging. If they meet the doctor's diagnostic and treatment requirements, the intermediate segmentation result of round 1 + the current round (i.e., round 2) is used as the target segmentation result; if they do not meet the doctor's diagnostic and treatment requirements, the number of rounds is increased, and rounds 1 + 2 + the current round (i.e., round 3) are selected for merging again to determine if they meet the doctor's requirements, and this process is repeated continuously. Alternatively, all possible merging processes can be performed, and then one of the merged results can be selected for the doctor's use.

[0251] For example, taking the first round of segmentation as the first neural network method, the second round as the second neural network method, and the third round as the region growing method, the current round, i.e. the third round, is the first type of segmentation method. Therefore, at least one of the first type of segmentation results and the second type of segmentation results obtained in the previous two rounds can be selected and merged with the first type of segmentation result obtained in the current round (i.e. the third round). Therefore, the intermediate segmentation results can include: 3+1, 3+2, 3+1+2.

[0252] For example, taking the first round of segmentation as the first neural network method, the second round as the region growing method, and the third round as the second neural network method, the current round, i.e. the third round, is the second type of segmentation method. Therefore, at least one of the first type of segmentation results obtained in the previous two rounds can be selected and merged with the first type of segmentation result obtained in the current round (i.e. the third round). Therefore, the intermediate segmentation results can include: 3+1, 3+2, 3+1+2.

[0253] For example, taking a segmentation round of 1 using the first neural network method, the second round using the region growing method, the third round using the second neural network method, and the fourth round using the morphological segmentation method as an example, if the current round (i.e., the fourth round) uses the second type of segmentation method, then at least one of the first type of segmentation results obtained in the previous three rounds can be selected and merged with the first type of segmentation result obtained in the current round (i.e., the fourth round). Therefore, the segmentation results can include: 4+1, 4+2, 4+1+2. Alternatively, at least one of the first type of segmentation results obtained in the previous three rounds and at least one of the second type of segmentation results obtained in the previous three rounds can be selected and merged with the second type of segmentation result obtained in the current round (i.e., the fourth round). Therefore, the intermediate segmentation results can include: 4+1+3, 4+2+3, 4+1+2+3.

[0254] When multiple merged intermediate segmentation results meet the doctor's diagnostic and treatment requirements, one of them can be selected for the doctor's use. The selection criteria can be flexibly set. For example, the higher the level of bronchus included, the better; or the more complete the bronchus of a certain level that the doctor is concerned about, the better; or the doctor observes the merged results and selects one of the merged results as the target segmentation result; or it can be selected randomly.

[0255] The embodiments of this application flexibly merge the segmentation results obtained by different segmentation methods to obtain a segmentation result, thereby improving the segmentation accuracy and the segmentation effect, which is more conducive to doctors' diagnosis and treatment.

[0256] In another embodiment of the lung image segmentation method provided in this application, taking the first round of segmentation as a first neural network method, the second round as a region growing method, the third round as a second neural network method, and the fourth round as a morphological segmentation method as an example, the segmentation method includes:

[0257] S1: Acquire lung images and determine the type of lung disease based on the lung images.

[0258] S2: The disease type is the target disease type. Configure one of the first segmentation methods, namely the first neural network method, to segment the lung image and obtain the segmentation result.

[0259] S3: If the segmentation result does not include bronchus with a stratum level not lower than the first preset value, proceed to step S4;

[0260] S4: Configure one of the segmentation methods that has not yet been configured in the first and second categories of segmentation methods, such as region growing, to segment the lung image and obtain the segmentation result;

[0261] S5: Based on 1+2, determine the intermediate segmentation result;

[0262] S6: If 1+2 does not meet the doctor's diagnostic and treatment requirements, return to execute S4, that is, configure one of the segmentation methods that have not yet been configured in the first and second categories of segmentation methods, such as the second neural network method, to segment the lung image and obtain the segmentation result;

[0263] S5: Based on at least two of all segmentation results, such as 3+1, 3+2, 3+1+2, determine the intermediate segmentation result;

[0264] S6: If 3+1, 3+2, and 3+1+2 do not meet the doctor's diagnostic and treatment requirements, return to execute S4, that is, configure one of the segmentation methods that have not yet been configured in the first and second categories of segmentation methods, such as morphological segmentation, to segment the lung image and obtain the segmentation result;

[0265] S5: Based on at least two of all segmentation results, such as 4+1, 4+2, 4+1+2, 4+1+3, 4+2+3, 4+1+2+3, determine the intermediate segmentation result;

[0266] S6: Determine whether 4+1, 4+2, 4+1+2, 4+1+3, 4+2+3, and 4+1+2+3 meet the doctor's diagnostic and treatment requirements, such as including bronchi with a layer level not lower than the first preset value. If the requirements are met, select one as the target segmentation result.

[0267] For example, taking the first round of segmentation as the first neural network method, the second round as the second neural network method, and the third round as the region growing method, the segmentation result of the first round is as follows: Figure 3a As shown, the results of the second round of segmentation are as follows: Figure 3b As shown, the partitioning result after merging 1 and 2 is as follows: Figure 3c As shown, the segmentation results of the third round are as follows: Figure 3d As shown, the segmentation results for rounds 1+2+3 are as follows: Figure 3e As shown.

[0268] In another embodiment of the lung image segmentation method provided in this application, the second type of segmentation method includes morphological segmentation, which is configured to segment the lung image to obtain a segmentation result, including:

[0269] After binarizing the lung image, an opening operation is performed, and connected components with a preset area are selected to obtain the corresponding high-level binarized image of the bronchus as the segmentation result. Specifically, this includes:

[0270] The lung images undergo initial denoising processing.

[0271] Based on the voxel intensity values ​​and a first higher-order threshold in the lung image, the lung image after the first denoising process is binarized to obtain a binarized image img_a, as shown below. Figure 4a As shown in the figure. In the binarized image img_a, voxels with intensity values ​​less than the higher-order threshold include bronchi, while voxels with intensity values ​​greater than the higher-order threshold do not include bronchi. For example, voxels with intensity values ​​less than the first higher-order threshold are marked as white, and voxels with intensity values ​​greater than the first higher-order threshold are marked as black.

[0272] Connectivity analysis is performed on the voxel portion of the binarized image img_a that does not include the bronchi. Connectivity components smaller than the first connectivity threshold are removed to obtain the binarized image img_b, as shown below. Figure 4b As shown;

[0273] An opening operation is performed on the binarized image img_b to obtain a binarized image img_c. The opening operation breaks the connection between the bronchi and the lung parenchyma in cases where there is a slight connection, such as an approximately gourd-shaped connection. This is done by filling the white area with black to indicate the disconnection, thereby separating the small bronchi.

[0274] Perform connected component analysis on the binarized graph img_c, removing connected components smaller than the second connected component threshold to obtain the binarized graph img_d, as shown below. Figure 4c As shown;

[0275] XORing the binarized images img_c and img_d yields the binarized image img_e corresponding to the higher-level bronchi, as shown below. Figure 4d As shown.

[0276] Optionally, the above operation can be performed on multiple slices of the lung image to obtain a binarized image of the higher-level bronchi in each slice.

[0277] This application embodiment separates small bronchi by performing connected component analysis and opening operations, thereby realizing the extraction of high-level bronchi through morphological segmentation methods.

[0278] Optionally, a second denoising process is performed on the lung image, and then a lung mask is extracted from the lung image after the second denoising process, including:

[0279] Based on the voxel intensity values ​​and the second higher-order threshold in the lung image after the second denoising process, the denoised lung image is binarized to obtain the binarized image img_1, as shown below. Figure 5a As shown, the optional second denoising process is more powerful than the first denoising process; otherwise, if the first denoising process is too powerful, it may lead to edge weakening. Typically, the second threshold is slightly higher than the first threshold.

[0280] The first and last slices of the binarized image img_1 are assigned the foreground value, for example, the foreground value is assigned as 1 and marked as white. The background area in the middle, for example, the background value is assigned as 0 and marked as black, is thus formed into a closed space, resulting in a new binarized image img_2.

[0281] Perform hole-filling operation on the new binarized image img_2 to obtain img_3, as shown below. Figure 5b As shown. Optionally, before performing the hole-filling operation, img_2 can be merged with the lung image or the lung image after the second denoising process;

[0282] Perform an XOR operation between the binarized image img_3 after filling the hole and img_2 to obtain img_4;

[0283] Connectivity analysis was performed on img_4, and the largest connected region was selected as the lung region. The binarized graph corresponding to the largest connected region is denoted as img_5, as shown below. Figure 5c As shown, img_5 is the extracted lung mask.

[0284] This application embodiment forms a closed space in the middle background area and then fills the hole to extract the lung mask, which is more conducive to removing interference outside the lung mask and obtaining better segmentation results.

[0285] Furthermore, the binarized image img_e and img_5 are ANDed together to form the binarized image img_e′ corresponding to the higher-order bronchus. Compared with img_e, img_e′ removes pseudo-higher-order bronchus outside the lung region.

[0286] This application combines morphological segmentation methods with methods for extracting lung masks, which helps to remove pseudo-higher-level bronchi outside the lung mask and improves the accuracy of segmentation.

[0287] Furthermore, when the first or second type of neural network method is configured to segment lung images, before inputting the lung image into the neural network to obtain matrix data, the method further includes: extracting the region of interest of the lung image, and then inputting the lung image containing only the region of interest into the neural network.

[0288] Optionally, the lung mask can be extracted first using the aforementioned method, with the same underlying principle, which will not be elaborated here. Then, the region of interest is determined based on the boundaries of the lung mask, for example, by constructing a rectangle based on the boundaries.

[0289] This application embodiment extracts the region of interest from the lung image and then inputs it into a neural network, which helps to remove pseudo-high-level bronchi outside the lung mask and improves the accuracy of segmentation.

[0290] Optionally, determine whether the lung image is from an animal or a human based on its features, including:

[0291] Extracting the lung mask;

[0292] Based on the lung mask, the first length of the 0th-order bronchus and the second length of the 0th-order bronchus to the bottom of the lung mask are obtained.

[0293] When the ratio of the first length to the second length is less than or equal to a preset ratio range, the lung image is an image of a human organ. When the ratio of the first length to the second length is less than or equal to a preset ratio range, the lung image is an image of an animal organ.

[0294] Optionally, the lung mask can be extracted first using the aforementioned method, with the same underlying principle, and will not be repeated here. This embodiment of the application uses the extracted lung mask to determine the features of the lung image to classify it as either an animal or a human, thereby facilitating the selection of a suitable segmentation method for the lung image.

[0295] It should be noted that other sorting schemes that can be easily conceived by those skilled in the art within the technical scope disclosed in this invention should also be within the protection scope of this invention, and will not be elaborated here.

[0296] Figure 6 This is a schematic diagram of image processing provided in an embodiment of the present invention. For example... Figure 6 As shown, the terminal device 6 in this embodiment includes: a processor 60, a memory 61, and a computer program 62 stored in the memory 61 and executable on the processor 60. When the processor 60 executes the computer program 62, it implements the steps in the above-described health quantification assessment method or product information labeling method embodiments, for example, as... Figure 1 Steps S1 to S6 shown Figure 2 Steps 201 to 207 are shown.

[0297] For example, computer program 62 can be divided into one or more modules / units, one or more of which are stored in memory 61 and executed by processor 60 to complete the present invention. One or more modules / units can be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of computer program 62 in terminal device 6.

[0298] Terminal device 6 can be a computing device such as a desktop computer, laptop, handheld computer, or cloud server. The terminal device may include, but is not limited to, processor 60 and memory 61. Those skilled in the art will understand that... Figure 6 This is merely an example of terminal device 6 and does not constitute a limitation on terminal device 6. It may include more or fewer components than shown, or combine certain components, or different components. For example, terminal device may also include input / output devices, network access devices, buses, etc.

[0299] The processor 60 may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0300] The memory 61 can be an internal storage unit of the terminal device 6, such as a hard disk or RAM of the terminal device 6. The memory 61 can also be an external storage device of the terminal device 6, such as a plug-in hard disk, Smart Media Card (SMC), Secure Digital (SD) card, or Flash Card equipped on the terminal device 6. Furthermore, the memory 61 can include both internal and external storage units of the terminal device 6. The memory 61 is used to store computer programs and other programs and data required by the terminal device. The memory 61 can also be used to temporarily store data that has been output or will be output.

[0301] This application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps described in the various method embodiments above.

[0302] This application provides a computer program product that, when run on a mobile terminal, enables the mobile terminal to implement the steps described in the above-described method embodiments.

[0303] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this invention. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0304] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0305] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0306] In the embodiments provided by this invention, it should be understood that the disclosed apparatus / terminal devices and methods can be implemented in other ways. For example, the apparatus / terminal device embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0307] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0308] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0309] If an integrated module / unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable medium does not include electrical carrier signals and telecommunication signals.

[0310] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0311] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0312] As used in this application specification and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrase "if determination" or "if the described condition or event is detected" may be interpreted, depending on the context, as "once determination," "in response to determination," "once the described condition or event is detected," or "in response to the detection of the described condition or event."

[0313] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0314] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0315] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for segmenting lung images, characterized in that, include: S1: Acquire lung images; S2: Configure one of the first type of segmentation methods to segment the lung image and obtain the segmentation result; S3: Based on the lung image and the segmentation results, determine the relative positional relationship between the lesion and the reference point; S4: If the relative positional relationship meets the preset conditions, determine whether the relative distance between the bronchus and the lesion in the segmentation result is greater than the first distance threshold, or whether the segmentation result includes bronchus with a layer level not lower than the first preset value. If the relative distance between the bronchus and the lesion in the segmentation result is greater than the first distance threshold, or the segmentation result does not include bronchus with a layer level not lower than the first preset value, proceed to step S5. S5: Select one of the segmentation methods that have not yet been configured from the first and second type of segmentation methods to segment the lung image and obtain the segmentation result; S6: Determine the intermediate segmentation result based on at least two of all segmentation results; S7: If the relative distance between the bronchus and the lesion in the intermediate segmentation result is greater than the first distance threshold, or if the intermediate segmentation result does not include bronchus with a layer level not lower than the first preset value, return to execute S5; Otherwise, the intermediate segmentation result is used as the target segmentation result.

2. The segmentation method as described in claim 1, characterized in that, The relative positional relationship satisfies preset conditions, including: The relative distance between the lesion and the reference point is greater than the second distance threshold; or, The bronchus at which the lesion is located is at a level greater than a second preset value, and the second preset value is less than a first preset value.

3. The segmentation method as described in claim 1, characterized in that, Determining the relative positional relationship between the lesion and the reference point includes: The location of the lesion was determined based on the lung images; The location of the reference point is determined based on the segmentation results; Based on the location of the lesion and the location of the reference point, determine the relative distance between the lesion and the reference point or the bronchial level where the lesion is located.

4. The segmentation method as described in claim 1, characterized in that, Determining the relative positional relationship between the lesion and the reference point includes: The location of the lesion was determined based on the lung images; The skeleton lines are determined based on the segmentation results; Based on the location of the lesion and the skeletal lines, the position of the lesion within the human anatomical structure is determined; Based on the location of the lesion in the human anatomical structure, the bronchial level where the lesion is located is determined.

5. The segmentation method as described in claim 1, characterized in that, The first type of segmentation method is configured to segment the lung image to obtain the segmentation result of the first type of segmentation result, and the second type of segmentation method is configured to segment the lung image to obtain the segmentation result of the second type of segmentation result; The step of determining intermediate segmentation results based on at least two of all segmentation results includes: When the segmentation method configured in the current round is the first type of segmentation method, at least one of the first type of segmentation results and the second type of segmentation results obtained in all previous rounds of segmentation is selected and merged with the first type of segmentation result obtained in the current round of segmentation to obtain the intermediate segmentation result.

6. The segmentation method as described in claim 1, characterized in that, The first type of segmentation method is configured to segment the lung image to obtain the segmentation result of the first type of segmentation result, and the second type of segmentation method is configured to segment the lung image to obtain the segmentation result of the second type of segmentation result; The step of determining intermediate segmentation results based on at least two of all segmentation results includes: When the segmentation method configured in the current round is the second type of segmentation method, at least one of the first type of segmentation results obtained in all previous rounds of segmentation is selected and merged with the second type of segmentation result obtained in the current round to obtain the intermediate segmentation result.

7. The segmentation method as described in claim 1, characterized in that, The first type of segmentation method includes a first neural network method, and the second type of segmentation method includes a second neural network method. The first or second type of neural network method is configured to segment the lung image to obtain a segmentation result, including: The lung image is input into the neural network to obtain matrix data, which includes at least one voxel and the confidence level corresponding to the voxel. Voxels with confidence levels higher than the first threshold are identified as bronchi, and a target binarized image is obtained as the segmentation result.

8. The segmentation method as described in claim 7, characterized in that, The method further includes: If the relative distance between the bronchus and the lesion in the intermediate segmentation result is greater than the first distance threshold, or if the intermediate segmentation result does not include the bronchus with a layer level not lower than the first preset value, and there is no segmentation method that has not yet been configured in the first and second type of segmentation methods, the matrix data obtained by configuring the first or second neural network method to segment the lung image is obtained. Lower the first threshold to the second threshold, and identify voxels with confidence levels below the second threshold as bronchi to obtain a new binarized image; The new binarized image is XORed with the target binarized image to obtain the XORed binarized image. Perform connected component analysis on the binarized graph after XOR to obtain a binarized graph corresponding to multiple connected components; Select one candidate connected component from the plurality of connected components; The binarized graphs corresponding to the candidate connected components are merged with the target binarized graph to obtain the merged binarized graph; When the number of connected components in the merged binarized graph is less than the number of connected components in the target binarized graph, the voxels in the candidate connected components are identified as bronchi, and the candidate connected components are identified as valid connected components. The binarized graph of the effective connected region is merged with the target binarized graph to obtain the segmentation result after connecting the broken bronchus.

9. The segmentation method as described in claim 1, characterized in that, The first type of segmentation method includes region growing, which is configured to segment the lung image to obtain a segmentation result, including: Seed points are selected in the lung images as the starting points for growth; Within the neighborhood of the seed point, at least one voxel with an intensity value less than the third threshold is selected; The initial value of the dynamic threshold is determined based on the intensity value of the seed point and the intensity value of the selected at least one voxel. Region growth is performed based on the seed points, and voxels with initial values ​​less than the dynamic threshold are selected as bronchi to obtain a binarized image as the segmentation result.

10. The segmentation method as described in claim 9, characterized in that, The method includes: Region growth is performed based on the initial value of the dynamic threshold, and voxels with intensity values ​​less than the initial value of the dynamic threshold are identified as new seed voxels and designated as bronchi. A new threshold is determined based on the initial value of the dynamic threshold and the intensity value of the new seed voxel when the number of new seed voxels reaches a preset number at fixed intervals. The new threshold is set as the initial value of the dynamic threshold, and the return step performs region growth based on the initial value of the dynamic threshold.

11. The segmentation method as described in claim 1, characterized in that, The second type of segmentation method includes morphological segmentation, which is configured to segment the lung image to obtain segmentation results, including: The lung image was denoised. Based on the intensity values ​​and higher-order thresholds of voxels in the lung image, the denoised lung image is binarized to obtain a binarized image img_a. In the binarized image img_a, voxels with intensity values ​​less than the higher-order threshold include bronchi, while voxels with intensity values ​​greater than the higher-order threshold do not include bronchi. Connectivity analysis is performed on the voxel portion of the binarized image img_a that does not include the bronchi, and connected components smaller than the first connected component threshold are removed to obtain the binarized image img_b. Perform an opening operation on the binarized image img_b to obtain the binarized image img_c; Perform connected component analysis on the binarized graph img_c, and remove connected components that are less than the second connected component threshold to obtain the binarized graph img_d; XORing the binarized image img_c and the binarized image img_d yields a higher-level binarized image corresponding to the bronchus.

12. An image processing apparatus, characterized in that, The system includes a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor, when executing the computer program, implements the lung image segmentation method as described in any one of claims 1 to 11.

13. A surgical robot, characterized in that, include Main control panel; From operating the equipment; and An image processing apparatus, coupled to the main operating console and the slave operating device, and configured to perform the lung image segmentation method as described in any one of claims 1-11.

14. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the lung image segmentation method as described in any one of claims 1 to 11.

Images