CN111260669A

CN111260669A - Lung lobe segmentation method and device based on CT image

Info

Publication number: CN111260669A
Application number: CN202010097533.1A
Authority: CN
Inventors: 王瑜; 赵朝炜; 周越; 孙岩峰; 邹彤; 张欢; 刘丰恺; 黄秋峰; 李新阳; 王少康; 陈宽
Original assignee: Infervision Co Ltd
Current assignee: Beijing Infervision Technology Co Ltd; Infervision Co Ltd
Priority date: 2020-02-17
Filing date: 2020-02-17
Publication date: 2020-06-09

Abstract

The invention provides a lung lobe segmentation method and device based on a CT image, electronic equipment and a storage medium, and solves the problem of how to perform lung lobe segmentation. The lung lobe segmentation method comprises the following steps: extracting a group of lung region images from the CT image sequence; identifying a plurality of lung fissure keypoints in the set of lung region images using a first neural network model; and carrying out lung lobe segmentation on the group of lung region images according to the plurality of lung fissure key points to obtain a group of lung lobe segmentation images.

Description

Lung lobe segmentation method and device based on CT image

Technical Field

The invention relates to the technical field of computer-aided medicine, in particular to a lung lobe segmentation method and device based on a CT image, electronic equipment and a storage medium.

Background

In recent years, with the rapid development of modern society and the increasing number of industrial activities, air pollution is more serious, and the incidence rate of lung diseases is also increased year by year. Therefore, analyzing the lung composition structure and further exploring the cause and development of the lung disease has important clinical significance for the diagnosis and treatment of the lung disease, wherein the identification of lung lobes is one of the key research directions for analyzing the lung composition structure.

Disclosure of Invention

In view of the above, embodiments of the present invention are directed to a method and an apparatus for lung lobe segmentation based on a CT image, and an electronic device and a storage medium to solve the problem of how to perform lung lobe segmentation.

The invention provides a lung lobe segmentation method based on a CT image, which comprises the following steps: extracting a group of lung region images from the CT image sequence; identifying a plurality of lung fissure key points in a set of lung region images by using a first neural network model; and carrying out lung lobe segmentation on the group of lung region images according to the plurality of lung fissure key points to obtain a group of lung lobe segmentation images.

In one embodiment, extracting a set of images of the lung region from the sequence of CT images includes: respectively extracting rib regions and rough lung segmentation regions from the CT image sequence; taking the rib region as a boundary, and expanding the rough lung segmentation region to the boundary outwards according to a preset step length to obtain a fine lung segmentation region; and cutting the fine lung segmentation areas in the CT image sequence to obtain a group of lung area images.

In one embodiment, extracting the rib region from the CT image sequence includes: segmenting a bone region from the CT image sequence based on the CT value of the bone; the rib region is identified from the bone region based on structural characteristics of the rib.

In one embodiment, segmenting the bone region from the CT image sequence based on the CT values of the bone comprises: and setting a first CT value threshold value based on the CT value of the bone, and acquiring a connected region of which the CT value is greater than or equal to the first CT value threshold value in the CT image as the bone region.

In one embodiment, after acquiring a connected region in the CT image with a CT value greater than or equal to the first CT value threshold, the method further includes: and removing the area of the connected area which is smaller than the preset area threshold value.

In one embodiment, after acquiring a connected region in the CT image with a CT value greater than or equal to the first CT value threshold, the method further includes: connected regions located within the coarsely segmented image are removed.

In one embodiment, identifying the rib region from the bone region based on structural characteristics of the rib comprises: and performing multi-plane reconstruction on the CT image sequence to obtain a reconstructed view, and identifying a rib region from the reconstructed view.

In one embodiment, identifying the rib region from the bone region based on structural characteristics of the rib comprises: and selecting the bone regions with the radian within a preset radian range and the widths of gaps between the adjacent bone regions within a preset width range in the bone region image as rib regions.

In one embodiment, extracting the coarse lung segmentation region from the CT image sequence comprises: and identifying a coarse lung segmentation region from the CT image sequence by using a second neural network model.

In one embodiment, before extracting the rib region and the coarse lung segmentation region from the CT image sequence, respectively, the method further includes: preprocessing the CT image sequence, the preprocessing comprising any one or combination of more of the following operations: removing background, removing white noise, cutting image, and transforming window width and level.

In one embodiment, further comprising: training a first neural network model; training the first neural network model comprises: acquiring a plurality of CT images including lung regions marked with lung fissure key points; and training the first neural network model by taking a plurality of CT images comprising lung regions as a training set to obtain the trained first neural network model.

In one embodiment, training the first neural network model further comprises: adding images with segmentation effects lower than a preset satisfaction degree in a group of lung lobe segmentation images into a training set to obtain an updated training set; and optimizing the first neural network model by using the updated training set.

In one embodiment, segmenting lung lobes of a set of images of the lung region according to a plurality of fissure key points, and obtaining a set of segmented lung lobes images comprises: constructing a fissure curved surface according to the multiple fissure key points; and carrying out lung lobe segmentation on the set of lung region images by using the lung fissure curved surface to obtain a set of lung lobe segmentation images.

In one embodiment, constructing the fissured surface from the plurality of fissured keypoints comprises: and performing surface fitting on the plurality of key points of the fissure of lung by using a surface fitting algorithm to obtain the fissure of lung curved surface.

In one embodiment, segmenting the set of images of the lung region using the lung fissure curve to obtain a set of segmented images of the lung lobes comprises: and taking the fissured lung curved surface as a foreground, performing distance transformation on the group of lung region images to divide the right lung into three lung lobes and the left lung into two lung lobes to obtain a group of lung lobe segmentation images.

The second aspect of the present invention provides a lung lobe segmentation apparatus based on CT image, comprising: the extraction module extracts a group of lung region images from the CT image sequence; an identification module that identifies a plurality of lung fissure keypoints in a set of lung regions using a first neural network model; and the segmentation module is used for carrying out lung lobe segmentation on the group of lung region images according to the plurality of lung fissure key points to obtain a group of lung lobe segmentation images.

A third aspect of the present invention provides an electronic device, comprising a memory, a processor and a computer program stored on the memory and executed by the processor, wherein the processor implements the steps of the CT image-based lung lobe segmentation method according to any one of the above methods when executing the computer program.

A fourth aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when being executed by a processor, implements the steps of the CT image-based lung lobe segmentation method according to any one of the above.

According to the lung lobe segmentation method and device based on the CT image, the electronic device and the storage medium, the first neural network model is utilized, the plurality of lung lobe key points are identified from the group of lung region images extracted from the CT image sequence, the lung lobe segmentation is carried out on the group of lung region images according to the plurality of lung lobe key points, the lung can be finely segmented into five lung lobes, and therefore a more accurate data basis is provided for clinical diagnosis.

Drawings

Fig. 1 is a flowchart illustrating a lung lobe segmentation method based on CT images according to a first embodiment of the present disclosure.

Fig. 2 is a flowchart illustrating a lung lobe segmentation method based on CT images according to a second embodiment of the present disclosure.

Fig. 3 is a flowchart illustrating a lung lobe segmentation method based on CT images according to a third embodiment of the present disclosure.

Fig. 4 is a flowchart illustrating a lung lobe segmentation method based on CT images according to a fourth embodiment of the present disclosure.

Fig. 5 is a schematic structural diagram of a lung lobe segmentation apparatus based on a CT image according to an exemplary embodiment of the present application.

FIG. 6 illustrates a block diagram of an electronic device in accordance with an embodiment of the present application.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

Summary of the application

CT image detection is one of the most important and accurate methods for determining whether a subject is a pneumonia patient, and recently, clinical examination mainly using CT image detection is determined as one of means for determining a novel coronavirus. In order to accurately locate the lesion position, it is generally necessary to segment the lung lobes after obtaining the CT image, and obtain a lung lobe segmented image corresponding to the CT image. However, the existing lung lobe segmentation schemes have respective defects, so that the segmentation result is poor and satisfactory.

For example, chinese patent application No. 201710546860.9 provides a lung lobe segmentation method based on CT images, which extracts lung lobe fissure points from a lung region by using a hessian matrix, and then constructs lung lobe fissure surfaces by using the lung lobe fissure points to distinguish lung lobes. In the method, the fissure points of the lung lobes are difficult to find under the condition that the density of the lung fissure is low or the density in the lung is increased due to inspiration. For another example, a chinese patent application No. 201910061682.X provides a three-dimensional lung lobe segmentation method based on CT images, which determines a pulmonary fissure by using a watershed transformation algorithm after obtaining a bronchus and a pulmonary trachea by using an image segmentation method. This method relies on the integrity of the bronchi and pulmonary vessels, making it difficult to determine the exact location of the lung fissure if there is a lesion in the bronchi or pulmonary vessels. For another example, chinese patent application No. 201811425317.4 provides a lung lobe segmentation method based on a full convolution neural network, which trains the full convolution neural network based on artificially labeled lung lobes as a sample set. In some cases, the boundaries between lobes are not clear, such as lobes with symptoms of pneumonia, making the model difficult to converge.

In view of the above, the present application provides a method and an apparatus for segmenting lung lobes based on CT images, a computer device and a storage medium, which identify a plurality of lung fissure key points in a set of lung region images by using a first neural network model; and carrying out lung lobe segmentation on the group of lung region images according to the plurality of lung fissure key points to obtain a group of lung lobe segmentation images. According to the scheme, the lung lobe segmentation of the CT image is realized, and meanwhile, the defects of the existing lung lobe segmentation method can be overcome.

Exemplary method

Fig. 1 is a flowchart illustrating a lung lobe segmentation method based on CT images according to a first embodiment of the present disclosure. The method may be used for a server or a terminal medical device. As shown in fig. 1, the lung lobe segmentation method 100 includes the following steps:

step S110, a set of lung region images is extracted from the CT image sequence.

The sequence of CT images includes a plurality of images at a plurality of CT slices. For a CT image sequence of the lungs, a three-dimensional reconstruction of the CT image sequence may approximately reveal a three-dimensional image of the lungs.

The lung region image refers to an image including only the lungs, for example, an image in which the lungs are black and other regions are white. A group of lung region images correspond to the CT image sequence one by one, namely, one lung region image is extracted from each CT image.

In step S120, a plurality of key points of the lung fissure in a set of images of the lung region are identified by using the first neural network model.

The lungs of the human body include five lobes, the left lung includes two lobes, and the right lung includes three lobes. The separation between the lobes of the lung is also called as fissures, which is an important mark for differentiating the lobes of the lung. The key points of the lung fissure are selected points from the lung fissure.

In one embodiment, the first neural network model is a centret model.

In one embodiment, the lung lobe segmentation method 100 further includes training a first neural network model.

The training process of the first neural network model specifically includes: acquiring a plurality of CT images including lung regions marked with lung fissure key points; and training the first neural network model by taking a plurality of CT images comprising lung regions as a training set to obtain the trained first neural network model.

The labeling mode can be manual labeling. The CT image comprising the lung region may be a chest CT image or a lung region image extracted from a sequence of CT images.

In one embodiment, the training process of the first neural network model further comprises: adding images with segmentation effects lower than a preset satisfaction degree in a group of lung lobe segmentation images into the training set to obtain an updated training set; and optimizing the first neural network model by using the updated training set. This results in a robust first neural network model.

And S130, carrying out lung lobe segmentation on the group of lung region images according to the plurality of lung fissure key points to obtain a group of lung lobe segmentation images.

A lobe segmented image refers to an image that includes only the lungs and has well-defined segmentation boundaries between different lobes. For example, the three lobes of the right lung are red, green, and blue, respectively, the two lobes of the left lung are yellow and purple, respectively, and the other parts are black.

According to the lung lobe segmentation method based on the CT image, the lung can be segmented into five lung lobes in a refined manner, so that a more accurate data base is provided for clinical diagnosis. Meanwhile, the segmentation method overcomes the defects of the existing lung lobe segmentation method, and is wider in application range and higher in detection precision.

Fig. 2 is a flowchart illustrating a lung lobe segmentation method based on CT images according to a second embodiment of the present disclosure. The difference between the segmentation method 200 and the segmentation method 100 shown in fig. 1 is only that in the segmentation method 200, the step S110 is specifically performed as:

step S111 extracts a rib region from the CT image sequence. Specifically, step S111 includes:

first, based on the CT values of the bone, a bone region is segmented from the CT image sequence.

Since different tissue structures in the human body differ slightly in density, the linear absorption coefficients for X-rays differ for different tissue structures. For ease of calculation and discussion, Hounsfield divides the linear attenuation coefficient into 2000 units, called CT values, which are typically above 150HU for bones.

Based on this, in an embodiment, a first CT value threshold may be set based on the CT value of the bone, and a connected region in the CT image having a CT value greater than or equal to the first CT value threshold is acquired as the bone region image. And acquiring a connected region of which the CT value is greater than or equal to the first CT value threshold value in the CT image by setting the first CT value threshold value, so as to obtain the bone region image.

In an embodiment, after acquiring a connected region in the CT image whose CT value is greater than or equal to the first CT value threshold, the method may further include: and removing the area of the connected area which is smaller than a preset area threshold value. When the first CT threshold is set too large, a part of the bone region may be missed, and when the first CT threshold is set too small, due to the large CT value of the calcifications, certain calcifications may exist in the lung or heart region to become interference noise in the image of the bone region, and therefore, the calcifications need to be removed. Generally, the area of the calcified points is small, so that the interference of the calcified points on the image of the bone region can be eliminated by removing the area of the communicating area which is smaller than the preset area threshold value, wherein the area threshold value can be preset according to the practical application.

In an embodiment, after acquiring a connected region in the CT image whose CT value is greater than or equal to the first CT value threshold, the method may further include: connected regions located within the coarsely segmented image are removed. By removing the communication region in the roughly segmented image, the influence of calcification points in the lung or heart region on the final segmentation result can be avoided, and the subsequent segmentation precision can be improved.

Second, the rib region is identified from the bone region based on structural characteristics of the rib.

Since the ribs are regularly arranged relative to other bones, the ribs are usually regularly arranged outside the lungs and have a certain arc shape and a left-right symmetrical structure, so that the ribs can be easily identified from the bone region.

In one embodiment, a multi-planar reconstruction of a CT image sequence results in reconstructed views from which rib regions are identified. The rib can be distinguished from other bones in a three-dimensional view through multi-plane reconstruction, particularly in the sagittal position, and the characteristics of the rib, namely the arrangement and the shape of the rib, can be clearly known.

In one embodiment, the bone regions in the bone region image with the radian within the preset radian range and the widths of the gaps between the adjacent bone regions within the preset width range are selected as rib regions.

It should be understood that other ways of acquiring the rib region image may be selected according to the requirements of the practical application scenario, for example, acquiring the rib region image directly through a neural network model. The embodiment of the present application does not limit the specific manner of acquiring the rib region image.

In step S112, a coarse lung segmentation region is extracted from the CT image sequence.

In one embodiment, step S112 is specifically performed as: and identifying a coarse lung segmentation region from the CT image sequence by using a second neural network model. The second neural network model may be a Unet model. The training mode of the second neural network model can be as follows: the CT image of the lung region which has been identified and marked by medical professionals is selected as a training sample of the neural network model to train the neural network model. Because only the rough segmentation image of the lung region in the CT image is obtained in the step, but not the accurate segmentation image of the lung region, a proper number of training samples can be selected to train the neural network model, so that the efficiency of the whole lung segmentation is improved.

In another embodiment, step S112 is specifically performed as: and selecting a region with the CT value within the CT value range of the lung region as a coarse segmentation image according to the CT value of the lung region. The CT number is a measure of the density of a local tissue or organ in the human body, and is commonly called Hounsfield Unit (HU), wherein the CT number of air is-1000 and the CT number of dense bone is + 1000. In fact, the CT value is a corresponding value of each tissue in the CT image corresponding to the X-ray attenuation coefficient, and the CT value is not an absolutely invariant value, and is related to not only internal factors of the human body such as respiration and blood flow, but also external factors such as X-ray tube voltage, CT apparatus, indoor temperature, and the like. The CT values of other tissues except bones in human tissues are below 300 and above-80, wherein the CT value of calcifications is 80-300, and the CT value of fat is-20 to-80. Because the lung region is basically air and the CT value is lower than other tissues, the CT value range can be set, and the connected region with the CT value in the CT value range is selected as the rough lung segmentation region.

In one embodiment, step S112 further comprises: and carrying out corrosion operation on the rough segmentation area of the lung to obtain a corroded rough segmentation image. The etching operation is a morphological operation, and the specific operation process is as follows: pixels are removed along the boundary of the object in the image and the object is reduced in size, i.e., the boundary of the object is reduced to remove noise from the object in the image. Since the rough lung segmentation region is usually not an accurate lung region, for example, the rough lung segmentation region may include other regions outside the lung region, i.e., noise regions in the rough lung segmentation region, the noise interference regions in the rough lung segmentation region may be removed by erosion to ensure that the eroded rough lung segmentation region is a part of the lung region and does not include regions outside the lung region.

And S113, expanding the rough lung segmentation region to the boundary outwards according to a preset step length by taking the rib region as the boundary to obtain a fine lung segmentation region.

In an embodiment, the specific implementation manner of step S113 may be: and taking the rib region as a boundary and the rough lung segmentation region as an interested region, and segmenting the CT image by a preset step length through the active contour model to obtain a fine lung segmentation region. Since the lung is tightly wrapped by the rib, that is, the rib region is the outer boundary of the lung region, the roughly segmented image can be used as a seed region or an interested region, and the rib region image is used as a boundary, the CT image is segmented by the active contour model and the preset step length, that is, the CT image is expanded from the interested region to the periphery by the preset step length until the CT image is expanded to the rib region, so as to obtain an image of the accurately segmented lung region, thereby providing accurate basic image data for subsequent lung lobe segmentation, pneumonia judgment and the like. The preset step length can be adjusted according to actual requirements, and can be properly reduced for obtaining higher precision. In further embodiments, the active contour model may comprise a LevelSet model or a Snake model. It should be understood that different active contour models may be selected according to requirements of an actual application scenario in the embodiment of the present application, as long as the selected active contour model can obtain an accurate image of a lung region by using the rib region image as a boundary and the roughly segmented image as a seed region, and the specific structure of the active contour model is not limited in the embodiment of the present application.

In one embodiment, step S113 further includes: and smoothing the boundary of the fine lung segmentation region. Because only part of the boundary of the lung region can be determined according to the rib region and the rib region is expanded by the active contour model in the same preset step length, the boundary of the obtained fine lung segmentation region may not be smooth, and after the fine lung segmentation region is obtained, the boundary of the fine lung segmentation region is subjected to smoothing treatment, so that the more accurate fine lung segmentation region can be obtained.

And step S114, cutting the fine lung segmentation areas in the CT image sequence to obtain a group of lung area images.

According to the lung lobe segmentation method provided by the embodiment, the structure of the outer ribs of the lung of a pneumonia patient is not affected by diseases and is not changed, the thick segmentation areas of the lung and the rib areas wrapping the lung areas are obtained by respectively obtaining the thick segmentation areas of the lung and the rib areas wrapping the lung areas, and the precision of lung segmentation is improved by taking the rib areas as the boundaries of the lung areas.

Fig. 3 is a flowchart illustrating a lung lobe segmentation method based on CT images according to a third embodiment of the present disclosure. As shown in fig. 3, the lung segmentation method 300 only differs from the lung segmentation method 200 shown in fig. 2 in that, in the lung segmentation method 300, step S110 further includes, before step S111:

step S115, a CT image sequence is preprocessed.

In an embodiment, the pre-processing may include any one or combination of the following operations: removing background, removing white noise, cutting image, and transforming window width and level. The specific implementation manner of removing the background may be: through setting the CT value range, the connected region in the CT value range is obtained, only the connected region with the largest area in the connected region is reserved, and other regions are set as background regions, so that the interference of other regions is eliminated. The specific implementation manner of removing the white noise may be: white noise caused in the process of taking the CT image is removed through a Gaussian filter. The specific implementation manner of cutting the image may be: and removing the background, and only reserving an effective area to reduce the complexity of subsequent image processing. The specific implementation manner of changing the window width and the window level may be: the region of interest is emphasized by setting values of the window width and the window level, so that interference of the region of no interest on subsequent processing is avoided, in the embodiment of the application, the window level may be selected to be-500, and the window width may be 1500, although it should be understood that the setting values of the window level and the window width may be adjusted according to actual situations.

Background and other interference factors in the CT image are discharged through preprocessing, so that the complexity of subsequent steps can be effectively reduced, and the lung segmentation efficiency is improved.

Fig. 4 is a flowchart illustrating a lung lobe segmentation method based on CT images according to a fourth embodiment of the present disclosure. As shown in fig. 4, the lung lobe segmentation method 400 differs from the lung lobe segmentation method 100 shown in fig. 1 only in that, in the lung lobe segmentation method 400, the step S130 specifically includes:

and S131, constructing a fissure curved surface according to the multiple fissure key points.

In one embodiment, step S131 is specifically performed as: and performing surface fitting on the plurality of key points of the fissure of lung by using a surface fitting algorithm to obtain the fissure of lung curved surface. For example, the surface fitting is performed by the least square method.

If all the extracted key points of the fissure are used for directly constructing the fissure curved surface, the obtained fissure curved surface is possibly not obvious enough. In order to highlight the fissure curved surface, a part of the fissure key points need to be removed. Thus, in one embodiment, step S131 is specifically performed as: and topologically connecting the multiple lung fissure key points, and removing the lung fissure key points which do not conform to the topological structure to obtain the optimized lung fissure curved surface. For example, a fissure keypoint that does not conform to the topology may be, for example, a fissure keypoint that forms a cusp on a curved surface. In another embodiment, each of the fissures key points corresponds to a three-dimensional coordinate (x, y, z), and since x and y define a plane, the z values of the fissures key points on the plane are compared, the point with the largest z value is found as the final representative fissures key point on the plane, and the rest of the fissures key points are removed.

In step S132, a lung lobe segmentation is performed on the set of lung region images by using the fissure curve to obtain a set of lung lobe segmentation images.

In one embodiment, step S132 is performed as: and taking the fissured lung curved surface as a foreground, performing distance transformation on the group of lung region images to divide the right lung into three lung lobes and the left lung into two lung lobes to obtain a group of lung lobe segmentation images. Specifically, first, distance transformation is performed on a group of lung region images, only the fissured lung surface is reserved, and the fissured lung surface image with black fissured lung surface and white remainder is obtained. And secondly, carrying out image registration on the group of lung region images by using the lung fissure curved surface image to obtain a group of lung lobe segmentation images.

Since the fissured surfaces obtained in step S131 are not completely connected, the obtained fissured surfaces cannot be directly used to perform lung lobe segmentation on a set of lung region images. After distance transformation, a more accurate fissured lung curved surface can be obtained, and lung lobe segmentation is realized.

Exemplary devices

Fig. 5 is a schematic structural diagram of a lung lobe segmentation apparatus based on a CT image according to an exemplary embodiment of the present application. As shown in fig. 5, the lung lobe segmentation apparatus 50 includes an extraction module 51, an identification module 52 and a segmentation module 53. The extraction module 51 is configured to extract a set of images of lung regions from the CT image sequence. The identification module 52 is configured to identify a plurality of lung fissure keypoints in a set of lung regions using a first neural network model. The segmentation module 53 is configured to perform lung lobe segmentation on the set of lung region images according to the plurality of lung fissure key points, so as to obtain a set of lung lobe segmentation images.

In one embodiment, the extraction module 51 specifically includes an extraction sub-module 511, an expansion module 512, and a cropping module 513. The extraction sub-module 511 is used for extracting the rib region and the coarse lung segmentation region from the CT image sequence. The expansion module 512 is configured to expand the rough lung segmentation region to the boundary according to a preset step length by using the rib region as the boundary, so as to obtain a fine lung segmentation region. The cropping module 513 crops the fine lung segmentation areas in the CT image sequence to obtain a set of lung region images.

The extraction submodule 511 is specifically configured to segment a bone region from a CT image sequence based on CT values of the bone; and identifying a rib region from the bone region based on structural characteristics of the rib. And identifying a coarse lung segmentation region from the CT image sequence by using the second neural network model.

In one embodiment, the extraction module 51 further comprises a pre-processing module 514 for pre-processing the sequence of CT images. The pre-treatment comprises any one or combination of more of the following operations: removing background, removing white noise, cutting image, and transforming window width and level.

In one embodiment, the segmentation module 53 is specifically configured to construct a fissured surface from a plurality of fissured keypoints; and carrying out lung lobe segmentation on the set of lung region images by using the lung fissure curved surface to obtain a set of lung lobe segmentation images.

The lung lobe segmentation device based on the CT image provided by the embodiment of the present invention is the same as the lung lobe segmentation method based on the CT image provided by the embodiment of the present invention, and the lung lobe segmentation device based on the CT image provided by any embodiment of the present invention can be executed. For details of the technique not described in detail in this embodiment, reference may be made to the lung lobe segmentation method based on CT images provided in the embodiment of the present invention, and details are not repeated here.

Exemplary electronic device

Next, an electronic apparatus according to an embodiment of the present application is described with reference to fig. 6. The electronic device may be either or both of the first device and the second device, or a stand-alone device separate from them, which stand-alone device may communicate with the first device and the second device to receive the acquired input signals therefrom.

FIG. 6 illustrates a block diagram of an electronic device in accordance with an embodiment of the present application. As shown in fig. 6, the electronic device 60 includes one or more processors 61 and a memory 62.

The processor 61 may be a Central Processing Unit (CPU) or other form of processing unit having data processing capabilities and/or instruction execution capabilities, and may control other components in the electronic device 60 to perform desired functions.

Memory 62 may include one or more computer program products that may include various forms of computer-readable storage media, such as volatile memory and/or non-volatile memory. The volatile memory may include, for example, Random Access Memory (RAM), cache memory (cache), and/or the like. The non-volatile memory may include, for example, Read Only Memory (ROM), hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium and executed by the processor 61 to implement the above-described CT image-based lung page segmentation method according to various embodiments of the present application and/or other desired functions. Various contents such as an input signal, a signal component, a noise component, etc. may also be stored in the computer-readable storage medium.

In one example, the electronic device 60 may further include: an input device 63 and an output device 64, which are interconnected by a bus system and/or other form of connection mechanism (not shown).

For example, when the electronic device is a first device or a second device, the input means 63 may be a camera for capturing an input signal of an image. When the electronic device is a stand-alone device, the input means 63 may be a communication network connector for receiving the acquired input signals from the first device and the second device.

The input device 63 may also include, for example, a keyboard, a mouse, and the like.

The output device 64 may output various information including the determined distance information, direction information, and the like to the outside. The output devices 64 may include, for example, a display, speakers, a printer, and a communication network and remote output devices connected thereto, among others.

Of course, for the sake of simplicity, only some of the components of the electronic device 60 relevant to the present application are shown in fig. 6, and components such as buses, input/output interfaces, and the like are omitted. In addition, the electronic device 60 may include any other suitable components depending on the particular application.

Exemplary computer program product and computer-readable storage Medium

In addition to the above-described methods and apparatus, embodiments of the present application may also be a computer program product comprising computer program instructions which, when executed by a processor, cause the processor to perform the steps in the CT image based lung page segmentation method according to various embodiments of the present application described in the "exemplary methods" section above of this specification.

The computer program product may be written with program code for performing the operations of embodiments of the present application in any combination of one or more programming languages, including an object oriented programming language such as Java, C + + or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computing device, partly on the user's device, as a stand-alone software package, partly on the user's computing device and partly on a remote computing device, or entirely on the remote computing device or server.

Furthermore, embodiments of the present application may also be a computer-readable storage medium having stored thereon computer program instructions, which, when executed by a processor, cause the processor to perform the steps in the CT image based lung page segmentation method according to various embodiments of the present application described in the "exemplary methods" section above in this specification.

The computer-readable storage medium may take any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may include, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or a combination of any of the foregoing. More specific examples (a non-exhaustive list) of the readable storage medium include: an electrical connection having one or more wires, a portable disk, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

The foregoing describes the general principles of the present application in conjunction with specific embodiments, however, it is noted that the advantages, effects, etc. mentioned in the present application are merely examples and are not limiting, and they should not be considered essential to the various embodiments of the present application. Furthermore, the foregoing disclosure of specific details is for the purpose of illustration and description and is not intended to be limiting, since the foregoing disclosure is not intended to be exhaustive or to limit the disclosure to the precise details disclosed.

The block diagrams of devices, apparatuses, systems referred to in this application are only given as illustrative examples and are not intended to require or imply that the connections, arrangements, configurations, etc. must be made in the manner shown in the block diagrams. These devices, apparatuses, devices, systems may be connected, arranged, configured in any manner, as will be appreciated by those skilled in the art. Words such as "including," "comprising," "having," and the like are open-ended words that mean "including, but not limited to," and are used interchangeably therewith. The words "or" and "as used herein mean, and are used interchangeably with, the word" and/or, "unless the context clearly dictates otherwise. The word "such as" is used herein to mean, and is used interchangeably with, the phrase "such as but not limited to".

It should also be noted that in the devices, apparatuses, and methods of the present application, the components or steps may be decomposed and/or recombined. These decompositions and/or recombinations are to be considered as equivalents of the present application.

The previous description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other aspects without departing from the scope of the application. Thus, the present application is not intended to be limited to the aspects shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

The foregoing description has been presented for purposes of illustration and description. Furthermore, the description is not intended to limit embodiments of the application to the form disclosed herein. While a number of example aspects and embodiments have been discussed above, those of skill in the art will recognize certain variations, modifications, alterations, additions and sub-combinations thereof.

Claims

1. A lung lobe segmentation method based on CT images is characterized by comprising the following steps:

extracting a group of lung region images from the CT image sequence;

identifying a plurality of lung fissure keypoints in the set of lung region images using a first neural network model;

and carrying out lung lobe segmentation on the group of lung region images according to the plurality of lung fissure key points to obtain a group of lung lobe segmentation images.

2. The method of claim 1, wherein the extracting a set of lung region images from the CT image sequence comprises:

respectively extracting rib regions and rough lung segmentation regions from the CT image sequence;

expanding the rough lung segmentation region outwards to the boundary according to a preset step length by taking the rib region as the boundary to obtain a fine lung segmentation region;

and cutting the fine lung segmentation region in the CT image sequence to obtain the group of lung region images.

3. The method of claim 2, wherein the extracting the rib region from the CT image sequence comprises:

segmenting a bone region from the CT image sequence based on CT values of the bone;

a rib region is identified from the bone region based on structural characteristics of the rib.

4. The CT-image-based lung lobe segmentation method according to claim 3, wherein the bone-based CT values segmenting bone regions from the CT image sequence comprises:

setting a first CT value threshold value based on the CT value of the bone, and acquiring a connected region of which the CT value is greater than or equal to the first CT value threshold value in the CT image as the bone region.

5. The method for segmenting lung lobes based on CT images as claimed in claim 4, further comprising, after the acquiring of the connected regions in the CT image with CT values greater than or equal to the first CT value threshold:

and removing the region with the area smaller than the preset area threshold value in the communication region.

6. The method for segmenting lung lobes based on CT images as claimed in claim 4, further comprising, after the acquiring of the connected regions in the CT image with CT values greater than or equal to the first CT value threshold:

and removing the connected regions in the roughly segmented image.

7. The CT-image-based lung lobe segmentation method of claim 3, wherein the identifying a rib region from the bone region based on the structural characteristics of the ribs comprises:

and performing multi-plane reconstruction on the CT image sequence to obtain a reconstructed view, and identifying the rib region from the reconstructed view.

8. The CT-image-based lung lobe segmentation method of claim 3, wherein the identifying a rib region from the bone region based on the structural characteristics of the ribs comprises:

and selecting the bone regions with the radian within a preset radian range and the widths of gaps between the adjacent bone regions within a preset width range in the bone region image as the rib regions.

9. The method of claim 2, wherein the extracting the coarse lung segmentation region from the CT image sequence comprises:

and identifying the coarse lung segmentation region from the CT image sequence by using a second neural network model.

10. The method of claim 2, wherein before the extracting the rib region and the coarse lung segmentation region from the CT image sequence, respectively, the method further comprises:

pre-processing the sequence of CT images, the pre-processing comprising any one or combination of more of: removing background, removing white noise, cutting image, and transforming window width and level.

11. The method for segmenting lung lobes based on CT images as claimed in claim 1, further comprising: training the first neural network model; the training the first neural network model comprises:

acquiring a plurality of CT images including lung regions marked with lung fissure key points;

and training a first neural network model by taking the CT images comprising the lung region as a training set to obtain the trained first neural network model.

12. The method of CT image-based lung lobe segmentation as claimed in claim 11, wherein said training the first neural network model further comprises:

adding images with segmentation effects lower than a preset satisfaction degree in the group of lung lobe segmentation images into the training set to obtain an updated training set;

and optimizing the first neural network model by using the updated training set.

13. The method of claim 1, wherein the performing lung lobe segmentation on the set of images of the lung region according to the plurality of key points of lung fissure to obtain a set of images of lung lobe segmentation comprises:

constructing a fissure curved surface according to the plurality of fissure key points;

and carrying out lung lobe segmentation on the group of lung region images by using the fissure curved surface to obtain the group of lung lobe segmentation images.

14. The method of claim 13, wherein the constructing a fissured surface from the plurality of fissured keypoints comprises:

and performing surface fitting on the plurality of key points of the fissure so as to obtain the fissure curved surface by using a surface fitting algorithm.

15. The method of claim 13, wherein the performing lung lobe segmentation on the set of images of lung regions by using the fissured surfaces to obtain the set of images of lung lobe segmentation comprises:

and performing distance transformation on the group of lung region images by taking the fissured surfaces as the foreground so as to divide the right lung into three lobes and the left lung into two lobes to obtain the group of lung lobe segmentation images.

16. A lung lobe segmentation device based on CT images is characterized by comprising:

the extraction module extracts a group of lung region images from the CT image sequence;

an identification module that identifies a plurality of lung fissure keypoints in the set of lung regions using a first neural network model;

and the segmentation module is used for carrying out lung lobe segmentation on the group of lung region images according to the plurality of lung fissure key points to obtain a group of lung lobe segmentation images.

17. An electronic device comprising a memory, a processor and a computer program stored on the memory for execution by the processor, wherein the processor when executing the computer program implements the steps of the CT image based lung lobe segmentation method according to any one of claims 1 to 15.

18. A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the steps of the method for CT image based lung lobe segmentation as claimed in one of the claims 1 to 15.