CN113011509A

CN113011509A - Lung bronchus classification method and device, electronic equipment and storage medium

Info

Publication number: CN113011509A
Application number: CN202110321500.5A
Authority: CN
Inventors: 简伟健; 张欢; 王瑜; 陈宽; 王少康
Original assignee: Infervision Medical Technology Co Ltd
Current assignee: Infervision Medical Technology Co Ltd
Priority date: 2021-03-25
Filing date: 2021-03-25
Publication date: 2021-06-22
Anticipated expiration: 2041-03-25
Also published as: CN113011509B

Abstract

The embodiment of the invention discloses a method and a device for classifying lung bronchi, electronic equipment and a storage medium. Wherein, the method comprises the following steps: acquiring an initial lung image to be classified for a lung bronchus; performing bronchial segmentation on the lung bronchus based on a preset image segmentation algorithm to obtain a bronchial mask image; segmenting the bronchus of the lung in the bronchus mask image to obtain at least two bronchus segments; determining the classification result of the pulmonary bronchi based on the adjacency relation between the various bronchial segments and the segment characteristics of each bronchial segment. The technical scheme can effectively improve the precision and robustness of bronchus classification, and can achieve good classification technical effects on images of abnormal and pathological changes of bronchus.

Description

Lung bronchus classification method and device, electronic equipment and storage medium

Technical Field

The embodiment of the invention relates to an image processing technology, in particular to a method and a device for classifying lung bronchi, electronic equipment and a storage medium.

Background

The existing lung bronchus classification algorithm generally has two types: one method is to classify bronchi by a template matching method, the method firstly obtains a plurality of bronchi classified templates, and then matches the templates to the image data to be classified by a certain algorithm, but the method has lower robustness due to the fact that the bronchi have a lot of variations, especially the difference between the templates and the target image is large; the other method is to perform multi-classification segmentation on the bronchus directly through semantic segmentation, the method strongly depends on image quality, and when the segmentation effect of the bronchus is poor and branches are missing, a good classification result is difficult to obtain.

Disclosure of Invention

The embodiment of the invention provides a classification method and device of lung bronchi, electronic equipment and a storage medium, so as to realize accurate segmentation of coronary artery.

In a first aspect, an embodiment of the present invention provides a method for classifying lung bronchi, including:

acquiring an initial lung image to be classified for a lung bronchus;

performing bronchial segmentation on the lung bronchus based on a preset image segmentation algorithm to obtain a bronchial mask image;

segmenting the bronchus of the lung in the bronchus mask image to obtain at least two bronchus segments;

determining the classification result of the pulmonary bronchi based on the adjacency relation between the various bronchial segments and the segment characteristics of each bronchial segment.

In a second aspect, an embodiment of the present invention provides a classification apparatus for bronchi of lungs, including:

the image acquisition module is used for acquiring an initial lung image to be used for classifying the bronchial tubes of the lung;

the image segmentation module is used for performing bronchial segmentation on the lung bronchus based on a preset image segmentation algorithm to obtain a bronchial mask image;

the bronchus segmentation module is used for segmenting bronchus of the lung in the bronchus mask image to obtain at least two bronchus segments;

and the bronchial classification module is used for determining the classification result of the pulmonary bronchi based on the adjacency relation among the bronchial segments and the segment characteristics of each bronchial segment.

In a third aspect, an embodiment of the present invention further provides an electronic device, where the electronic device includes:

one or more processors;

a storage device for storing one or more programs,

when executed by the one or more processors, cause the one or more processors to implement a method of classifying pulmonary bronchi according to any of the embodiments of the present invention.

In a fourth aspect, the present invention further provides a computer-readable storage medium, on which a computer program is stored, which when executed by a processor implements the method for classifying pulmonary bronchi according to any one of the embodiments of the present invention.

According to the technical scheme of the embodiment of the invention, the bronchus mask image is obtained by performing bronchus segmentation on the obtained initial lung image, so that the interference information in the initial abdomen image can be removed, and a foundation is laid for further accurately classifying the bronchus of the lung; and finally, determining a classification result of the pulmonary bronchus based on the adjacency relation among the bronchial segments and the segment characteristics of each bronchial segment, and classifying the pulmonary bronchus through the topological structure and the characteristics of the pulmonary bronchus, so that the accuracy and the robustness of the classification of the bronchus can be effectively improved, and a good classification effect is achieved on images with abnormal and pathological changes of the bronchus.

Drawings

In order to more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, a brief description is given below of the drawings used in describing the embodiments. It should be clear that the described figures are only views of some of the embodiments of the invention to be described, not all, and that for a person skilled in the art, other figures can be derived from these figures without inventive effort.

FIG. 1 is a schematic flow chart of a method for classifying bronchi in lung according to an embodiment of the present invention;

fig. 2 is a flowchart illustrating a method for training a bronchial segment classification model according to an embodiment of the present invention;

fig. 3 is a schematic diagram of a training data sample of a bronchial segment classification model according to an embodiment of the present invention;

fig. 4 is a schematic diagram of an adjacency matrix constructed based on the adjacency relationship of the bronchial segments of the pulmonary bronchi according to an embodiment of the present invention;

FIG. 5 is a schematic flow chart diagram illustrating an alternative example of a method for classifying bronchi in a lung according to an embodiment of the present invention;

FIG. 6 is a schematic diagram of a testing process for classifying bronchus according to the method for classifying bronchus of lung according to the present invention;

FIG. 7 is a schematic structural diagram of a lung bronchus classifying device according to an embodiment of the present invention;

fig. 8 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.

Detailed Description

The present invention will be described in further detail with reference to the accompanying drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting of the invention. It should be further noted that, for the convenience of description, only some of the structures related to the present invention are shown in the drawings, not all of the structures.

It should be further noted that, for the convenience of description, only some but not all of the relevant aspects of the present invention are shown in the drawings. Before discussing exemplary embodiments in more detail, it should be noted that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although a flowchart may describe the operations (or steps) as a sequential process, many of the operations can be performed in parallel, concurrently or simultaneously. In addition, the order of the operations may be re-arranged. The process may be terminated when its operations are completed, but may have additional steps not included in the figure. The processes may correspond to methods, functions, procedures, subroutines, and the like.

Fig. 1 is a flowchart illustrating a bronchial classification method according to an embodiment of the present invention, where the method may be performed by a bronchial classification device, and the device may be configured in a terminal or a server, and the terminal and the server perform the bronchial classification method according to the embodiment of the present invention independently or cooperatively.

As shown in fig. 1, the bronchial classification method in this embodiment may specifically include:

s110, obtaining an initial lung image to be used for classifying the lung bronchus.

Illustratively, the initial lung image may be a lung image acquired based on a medical imaging device. The initial lung image may be an image that satisfies a DICOM (Digital Imaging and Communications in Medicine) protocol, among others. The initial lung image may be obtained from a medical imaging device in real time, from an image database, or from a lung image transmission received from an external device.

The bronchial classification may specifically be: after entering the thoracic cavity, the trachea divides into left and right main bronchi. The right main bronchus is divided into upper lobe bronchus and middle segment bronchus, the middle broken bronchus is divided into middle lobe bronchus and lower lobe bronchus, the left main bronchus is divided into upper lobe bronchus and lower lobe bronchus, the left upper lobe bronchus is divided into lingual segment bronchus branch, thus the right lung is divided into upper middle lower three lobes, and the left lung is divided into upper and lower two lobes. These bronchi are subdivided into segments, sub-segment bronchi, terminal bronchioles, respiratory bronchioles, alveolar ducts, alveolar sacs and alveoli.

It should be noted that the bronchus and the bronchus are taken as a whole in the embodiment of the present invention. The trachea is a component of a respiratory system, is connected with a pipeline between the larynx and the bronchus, the upper part of the trachea is connected with the larynx, and the lower part of the trachea is communicated with the left lung and the right lung. The bronchus is the first-stage bronchus divided from the trachea, namely the left and right main bronchus.

S120, performing bronchial segmentation on the lung bronchus based on a preset image segmentation algorithm to obtain a bronchial mask image.

Optionally, after obtaining an initial lung image to be classified into pulmonary bronchi, gray-scale normalization may be performed on the initial lung image, and then segmentation of the pulmonary bronchi is performed on the normalized initial lung image.

The preset image segmentation method may be various, and the embodiment of the present invention is not limited to which image segmentation method is specifically adopted. Illustratively, the bronchus of the lung can be segmented by a conventional image algorithm, such as a region growing algorithm and/or a Hessian matrix algorithm, and also can be segmented by an artificial intelligence method such as deep learning, for example, a 2D or 3D UNet deep learning segmentation network. It can be understood that when the lung bronchus is segmented, the more classes the better, the more classes the classification of the lung bronchus is helpful.

S130, segmenting the lung bronchus in the bronchus mask image to obtain at least two bronchus segments.

Specifically, a bronchial centerline of a pulmonary bronchus in the bronchial mask image may be extracted, and the pulmonary bronchus may be segmented based on the bronchial centerline to obtain at least two bronchial segments.

Among them, the bronchial centerline extraction can be understood as extracting a skeleton of the bronchial segmentation. Illustratively, the bronchial midline extraction may be performed using a fire simulation or maximal internal ball approach. The goal is to allow the bronchial midline to reflect the original bronchial topology. The bifurcation point of the bronchial midline is the bifurcation point of the original bronchus, and the bronchial midline is located at the center of each segment of bronchus as much as possible.

Optionally, the bronchial midline is the bronchial skeleton with a single pixel point as width. In other words, the bronchus centerline is a line formed by a plurality of continuous single pixel points located at the center of the bronchus along the extension direction of the bronchus.

Optionally, segmenting the pulmonary bronchus based on the bronchial midline comprises: extracting key points of the bronchial midline to obtain at least two target key points of the bronchial midline; segmenting the pulmonary bronchus based on the at least two target key points.

Specifically, extracting key points of the bronchial centerline to obtain target key points of the bronchial centerline may include: and determining a target key point of the bronchus central line based on the number of adjacent pixel points of each pixel point on the bronchus central line. For example, the target keypoint of the bronchial centerline may be determined based on whether the number of neighboring pixel points of each pixel point located on the bronchial centerline changes.

Wherein the target keypoints may include bifurcation points and end points. Aiming at each pixel point on the bronchus midline, if more than 3 pixel points are positioned on the bronchus midline in 26 neighborhoods of the pixel point, the pixel point is determined as a bifurcation point of the bronchus of the lung; and if only 1 pixel in the neighborhood is located at the upper point of the bronchial midline, determining the pixel point as the end point of the pulmonary bronchus. In order to avoid the influence of the condition of the bifurcation point causing many false detections on the classification of the pulmonary bronchus, redundant bifurcation points can be removed through an algorithm, for example, the distance between the bifurcation point and the bifurcation point can be calculated, and the bifurcation point with the distance within a preset distance range is subjected to de-duplication treatment, and the like.

S140, determining a classification result of the pulmonary bronchi based on the adjacency relation among the bronchial segments and the segment characteristics of each bronchial segment.

Optionally, determining adjacency relations among the bronchial segments, and constructing an adjacency matrix corresponding to the pulmonary bronchi based on the adjacency relations; determining a classification result of each of the bronchial segments based on the adjacency matrix and the segment characteristics of each of the bronchial segments; determining a classification result of the pulmonary bronchi based on the classification result of each of the bronchial segments.

Specifically, the determination of the adjacency relationship between the bronchial segments may be that the bronchial centerlines of the pulmonary bronchi are traversed by adopting a depth-first search method or a breadth-first search method, starting from the uppermost end point. Each time a target key point is passed, a bronchial segment is obtained. In other words, the bronchus between two adjacent key points may be one bronchial segment.

Illustratively, the adjacency between different segments may also be derived by whether two bronchial segments have the same target keypoints between them. When the whole bronchial midline of the pulmonary bronchus is traversed, the pixel points of all the bronchial segments and the adjacency relation among all the bronchial segments can be obtained.

Optionally, determining a classification result of each of the bronchial segments based on the adjacency matrix and the segment characteristics of each of the bronchial segments may include: inputting the adjacency matrix and the segment characteristics of each bronchial segment into a bronchial segment classification model which is trained in advance to obtain a classification result of each bronchial segment; the bronchial segment classification model is obtained by training a pre-established graph network model based on the adjacency relation among all bronchial segments of lung bronchi in a sample lung image and the segment characteristics of all the bronchial segments.

It should be noted that, the embodiment of the present invention does not limit the specific model structure or model form of the graph network model, and is within the protection scope of the embodiment of the present invention as long as the classification of the bronchial segments can be realized after training. Illustratively, the Graph Network model may be at least one of a Graph Convolutional neural Network (GCN) model, a Graph Attention Network (GAT) model, and a Graph Isomorphic Network (GIN) model.

As can be seen from the foregoing, the classification result of the bronchial segment may be a classification result of each bronchial segment on a medial line of a pulmonary bronchus, and in this case, the classification result of each bronchial segment on the medial line of the bronchus needs to be mapped onto the entire bronchus of the pulmonary bronchus, and the exemplary determination of the classification result of the pulmonary bronchus based on the classification result of each bronchial segment may include: determining a classification result of the pulmonary bronchi based on the classification result of each of the bronchial segments and a region growing algorithm. And growing all the segmented pixel points of the lung bronchus by region growing (region growing) operation aiming at the pixel points in each section of the bronchus section, and mapping the classification result of each pixel point on the middle line to the lung bronchus to obtain the final classification result of the lung bronchus.

Fig. 2 is a flowchart illustrating a method for training a bronchial segment classification model according to an embodiment of the present invention, and as shown in fig. 2, a specific implementation manner of the method for training the bronchial segment classification model may include:

step one, a training data set is obtained.

Fig. 3 is a schematic diagram of a training data sample of a bronchial segment classification model according to an embodiment of the present invention. As shown in fig. 3, the training data sample may include a bronchus mask image and a bronchus classified gold standard image, or a label image. The left image is a bronchus mask image (mask for short) obtained by performing bronchus segmentation on a sample lung image and used for inputting a neural network to be trained, and the right image is a gold standard image obtained by marking a bronchus classification on the bronchus mask image and used for calculating a loss function of the neural network, namely loss of the network.

And step two, establishing an initial graph network model.

Illustratively, the graph network model may be a GCN model. In order to make the GCN model accurate in classifying the bronchi of the lung, the GCN model in the embodiment of the present invention may be an operator without pooling layers or a posing or the like for changing the graph structure, alternatively, unlike the existing GCN model. I.e. each convolution, so that the topology of the graph remains consistent.

It should be noted that "one" and "two" of "step one" and "step two" are only used to distinguish different steps, and are not limited to the execution sequence. The first step and the second step can be executed in series or in parallel, and the execution sequence can be interchanged.

And step three, randomly sampling from the training data set, and performing data expansion on the samples.

In order to enable the classification effect after the GCN model training to be better and the classification result to be more accurate, a large amount of training data can be used for training the GCN model. However, considering the limitations of acquiring medical images and the limitation of training data, data expansion can be performed based on training samples in the training data set. Illustratively, the method of augmentation may include, but is not limited to, at least one of affine transformation, elastic deformation (e.g., stretching, clipping, etc.), adding noise, and pruning the airway tree.

And step four, extracting the bronchial midline.

That is, the mask of the bronchial segmentation is subjected to bronchial centerline extraction.

And step five, extracting bifurcation points and end points of the bronchial midline.

And step six, segmenting the lung bronchus according to the bifurcation point and the end point.

And step seven, calculating the adjacency relation among all the bronchial segments, and constructing an adjacency matrix corresponding to the lung bronchi based on the adjacency relation.

Fig. 4 is a schematic diagram of an adjacency matrix constructed based on the adjacency relationship of the bronchial segments of the pulmonary bronchi according to an embodiment of the present invention. As shown in fig. 4, the left side is to divide the bronchial tree of a pulmonary bronchus into different bronchial segments, and to mark each bronchial segment differently. And setting the values of the matrix subscripts corresponding to the two bronchial segments as 0 or 1 according to whether each two bronchial segments have an adjacency relation, namely whether the two bronchial segments are adjacent. For example, setting the value of the index position corresponding to two adjacent bronchial segments to 1, and setting the value of the index position corresponding to two non-adjacent bronchial segments to 0, may obtain the adjacency matrix shown on the right. If the bronchial segment 1 is adjacent to the bronchial segment 2, the matrix elements corresponding to the first row, the second column, and the second row, the first column are both 1; the

bronchial segments

1 and 3 are not adjacent, and the matrix elements corresponding to the third row and the third column and the third row and the first column are all 0. It is understood that the bronchial segment 1 and the bronchial segment 1 are the same bronchial segment and not in a neighboring relationship, so the matrix element corresponding to the first row and the first column is also 0. Here, the adjacency relation has no direction and no weight, that is, the adjacency matrix is a symmetric matrix having only 0 and 1 values.

And step eight, calculating the segment characteristics of each segment of the bronchial segment.

The segment characteristics may include, but are not limited to, a starting position, an ending position, a length, and a direction of the bronchial segment. Further, the segment characteristics of each segment of the bronchial segment may also be normalized.

Step nine, calculating the mode of the labels corresponding to each pixel point in each section of the bronchial segment, taking the mode as the section label, and executing the step eleven.

Specifically, the label on the gold standard image after the bronchial classification is labeled with the bronchial mask image (i.e., label shown on the right side of fig. 3) can be obtained by traversing each pixel point on the bronchial segment, and the label appearing most in the segment is counted as the label of the segment, so as to calculate the loss. That is, the largest number of pixel points corresponding to which label (category) in the bronchial segment is, the bronchial segment is considered to belong to which category.

It should be noted that "seven", "eight", and "nine" in "step seven", "step eight", and "step nine" are only for distinguishing different steps, and do not limit the execution order. The seventh step, the eighth step and the ninth step can be executed in series or in parallel, and the execution order can be interchanged.

And step ten, inputting the adjacency matrix and the segment characteristics of each bronchial segment into the GCN model for prediction to obtain a prediction result of each bronchial segment.

That is, the adjacency matrix in step seven and the segment features extracted in step eight are input into the GCN, and network forward propagation is performed to obtain the prediction result of each bronchial segment output by the network.

And step eleven, calculating the loss through the prediction result and the label, performing gradient back transmission, and updating the network parameters of the GCN model.

And step twelve, judging whether the GCN model is converged, if so, executing the step thirteen, otherwise, returning to execute the step three.

Specifically, whether the GCN model reaches the end training condition may be determined by determining whether the loss function loss of the GCN model converges. If the training ending condition is reached, executing a step thirteen; if the loss function does not converge, steps three through eleven may be repeatedly performed.

And step thirteen, finishing the training to obtain the bronchial segment classification model.

The embodiment of the invention provides a lung bronchus classification method based on a graph convolution network. Firstly, segmenting lung bronchus, extracting a bronchus central line from a mask obtained by segmentation, segmenting the bronchus central line, further extracting segment characteristics from a bronchus segment, calculating an adjacency matrix between all the bronchus segments, and then predicting the classification of the bronchus segment by taking the segment characteristics of the bronchus segment and the adjacency matrix as the input of a GCN (generalized computer network). In the embodiment of the invention, only the bronchial mask image is needed in the training process, and the initial lung image is not needed, so that the generalization capability of the bronchial segment classification model can be effectively improved.

Fig. 5 is a flowchart illustrating an alternative example of a method for classifying bronchi in lung according to an embodiment of the present invention. As shown in fig. 5, the method for classifying lung bronchi specifically includes:

first, an initial lung image of the lung to be classified for the bronchi of the lung is acquired. The initial lung image may be an image that satisfies a DICOM (Digital Imaging and Communications in Medicine) protocol, among others.

And then, performing bronchial segmentation on the lung bronchus based on a preset image segmentation algorithm to obtain a bronchial mask image.

Further, bronchial centerlines of the pulmonary bronchi in the bronchial mask image are extracted.

On the basis, key points can be extracted from the bronchial midline to obtain the bifurcation point and the end point of the bronchial midline.

And then, segmenting the pulmonary bronchus according to the bifurcation point and the end point to obtain all the bronchus segments of the pulmonary bronchus.

Furthermore, the adjacency relation between the bronchial segments is respectively calculated, an adjacency matrix is constructed based on the adjacency relation, and segment characteristics of each bronchial segment, such as the starting position and the ending position of the bronchial segment, the length of the bronchial segment, the extending direction of the bronchial segment, and the like, are calculated.

And then, inputting the adjacency matrix and the segment characteristics of each bronchial segment into a bronchial segment classification model which is trained in advance to obtain a classification result of each bronchial segment.

Finally, the classification result of the pulmonary bronchi is determined by a region growing algorithm based on the classification result of each bronchial segment.

According to the method provided by the embodiment of the invention, the accuracy and robustness of bronchus classification can be effectively improved by learning the topological structure and characteristics of the bronchus, and a good classification effect is achieved even for images of abnormal and pathological changes of the bronchus of the lung.

Fig. 6 is a schematic diagram of a testing process for classifying pulmonary bronchi according to the method for classifying pulmonary bronchi provided by the embodiment of the present invention. As shown in fig. 6, the testing process may specifically include:

1. an initial lung image, which may be, for example, a lung CT image, is acquired that is to classify the pulmonary bronchi. Then, the lung CT image can be subjected to gray level normalization and bronchial segmentation to obtain a bronchial mask image.

2. Extracting the bronchial centerline of the pulmonary bronchus in the bronchial mask image, segmenting the bronchial centerline to obtain bronchial segments, and calculating the end characteristics of the bronchial segments and the adjacency matrix of each bronchial segment.

3. And establishing a graph convolution neural network model, and loading the trained network weight, namely obtaining the bronchial segment classification model which is trained in advance.

4. Inputting the adjacency matrix and the characteristics of each segment into the bronchial segment classification model, and carrying out forward propagation.

5. The output of the network is operated on argmax to obtain the category of each bronchial segment.

6. And growing all pixels of the segmented lung bronchus by region growing operation aiming at the pixel point of each section of the bronchus section, and finally obtaining a lung bronchus grading result.

The lung bronchus grading method provided by the embodiment of the invention can effectively improve the precision and robustness of bronchus by learning the topological structure and characteristics of the lung bronchus.

Fig. 7 is a schematic structural diagram of a lung bronchus classifying device according to an embodiment of the present invention, which can be implemented by software and/or hardware, and includes: an image acquisition module 710, an image segmentation module 720, a bronchial segmentation module 730, and a bronchial classification module 740.

The image obtaining module 710 is configured to obtain an initial lung image of a lung to be classified; the image segmentation module 720 is configured to perform bronchial segmentation on the lung bronchus based on a preset image segmentation algorithm to obtain a bronchial mask image; a bronchus segmentation module 730, configured to segment a bronchus of a lung in the bronchus mask image to obtain at least two bronchus segments; a bronchial classification module 740 configured to determine a classification result of the pulmonary bronchi based on the adjacency relationship between the bronchial segments and the segment characteristics of each bronchial segment.

On the basis of any optional technical solution in the embodiment of the present invention, the bronchial classification module may include: the device comprises an adjacency matrix construction unit, a bronchial segment classification unit and a classification result determination unit.

The adjacency matrix construction unit is used for determining adjacency relations among all bronchial segments and constructing an adjacency matrix corresponding to the pulmonary bronchi based on the adjacency relations; a bronchial segment classification unit, configured to determine a classification result of each bronchial segment based on the adjacency matrix and a segment feature of each bronchial segment; a classification result determination unit that determines a classification result of the pulmonary bronchi based on the classification result of each of the bronchial segments.

On the basis of any optional technical solution in the embodiment of the present invention, the bronchial segment classification unit may be configured to:

inputting the adjacency matrix and the segment characteristics of each bronchial segment into a bronchial segment classification model which is trained in advance to obtain a classification result of each bronchial segment;

the bronchial segment classification model is obtained by training a pre-established graph network model based on the adjacency relation among all bronchial segments of lung bronchi in a sample lung image and the segment characteristics of all the bronchial segments.

On the basis of any optional technical solution in the embodiment of the present invention, the classification result determining unit may be configured to:

determining a classification result of the pulmonary bronchi based on the classification result of each of the bronchial segments and a region growing algorithm.

On the basis of any optional technical scheme in the embodiment of the present invention, the bronchial segment module may specifically be configured to:

and extracting the bronchial midline of the pulmonary bronchus in the bronchial mask image, and segmenting the pulmonary bronchus based on the bronchial midline to obtain at least two bronchial segments.

On the basis of any optional technical solution in the embodiment of the present invention, the bronchial segment module may include: a key point extraction unit and a bronchus segmentation unit.

The device comprises a key point extracting unit, a central processing unit and a central processing unit, wherein the key point extracting unit is used for extracting key points of the bronchial central line to obtain at least two target key points of the bronchial central line; segmenting the pulmonary bronchus based on the at least two target key points.

On the basis of any optional technical solution in the embodiment of the present invention, the keypoint extraction unit may be configured to:

and determining a target key point of the bronchus central line based on the number of adjacent pixel points of each pixel point on the bronchus central line.

The product can execute the bronchus classification method provided by any embodiment of the invention, and has corresponding functional modules and beneficial effects for executing the bronchus classification method.

Fig. 8 is a schematic structural diagram of an electronic device according to an embodiment of the present invention. FIG. 8 illustrates a block diagram of an exemplary electronic device 12 suitable for use in implementing embodiments of the present invention. The electronic device 12 shown in fig. 8 is only an example, and should not bring any limitation to the functions and the scope of use of the embodiment of the present invention.

As shown in FIG. 8, electronic device 12 is embodied in the form of a general purpose computing device. The components of electronic device 12 may include, but are not limited to: one or more processors or processing units 16, a system memory 28, and a bus 18 that couples various system components including the system memory 28 and the processing unit 16.

Bus 18 represents one or more of any of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, and a processor or local bus using any of a variety of bus architectures. By way of example, such architectures include, but are not limited to, Industry Standard Architecture (ISA) bus, micro-channel architecture (MAC) bus, enhanced ISA bus, Video Electronics Standards Association (VESA) local bus, and Peripheral Component Interconnect (PCI) bus.

Electronic device 12 typically includes a variety of computer system readable media. Such media may be any available media that is accessible by electronic device 12 and includes both volatile and nonvolatile media, removable and non-removable media.

The system memory 28 may include computer system readable media in the form of volatile memory, such as Random Access Memory (RAM)30 and/or cache memory 32. The electronic device 12 may further include other removable/non-removable, volatile/nonvolatile computer system storage media. By way of example only, storage system 34 may be used to read from and write to non-removable, nonvolatile magnetic media (not shown in FIG. 8, and commonly referred to as a "hard drive"). Although not shown in FIG. 7, a magnetic disk drive for reading from and writing to a removable, nonvolatile magnetic disk (e.g., a "floppy disk") and an optical disk drive for reading from or writing to a removable, nonvolatile optical disk (e.g., a CD-ROM, DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to bus 18 by one or more data media interfaces. System memory 28 may include at least one program product having a set (e.g., at least one) of program modules that are configured to carry out the functions of embodiments of the invention.

A program/utility 40 having a set (at least one) of program modules 42 may be stored, for example, in system memory 28, such program modules 42 including, but not limited to, an operating system, one or more application programs, other program modules, and program data, each of which examples or some combination thereof may comprise an implementation of a network environment. Program modules 42 generally carry out the functions and/or methodologies of the described embodiments of the invention.

Electronic device 12 may also communicate with one or more external devices 14 (e.g., keyboard, pointing device, display 24, etc.), with one or more devices that enable a user to interact with electronic device 12, and/or with any devices (e.g., network card, modem, etc.) that enable electronic device 12 to communicate with one or more other computing devices. Such communication may be through an input/output (I/O) interface 22. Also, the electronic device 12 may communicate with one or more networks (e.g., a Local Area Network (LAN), a Wide Area Network (WAN), and/or a public network, such as the Internet) via the network adapter 20. As shown in FIG. 8, the network adapter 20 communicates with the other modules of the electronic device 12 via the bus 18. It should be appreciated that although not shown in FIG. 8, other hardware and/or software modules may be used in conjunction with electronic device 12, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, among others.

The processing unit 16 executes programs stored in the system memory 28 to perform various functional applications and data processing, such as implementing a method for classifying pulmonary bronchi as provided by the present embodiment.

Embodiments of the present invention also provide a storage medium containing computer-executable instructions which, when executed by a computer processor, perform a method of classifying a pulmonary bronchus, the method comprising:

acquiring an initial lung image to be classified for a lung bronchus;

Computer storage media for embodiments of the invention may employ any combination of one or more computer-readable media. The computer readable medium may be a computer readable signal medium or a computer readable storage medium. A computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples (a non-exhaustive list) of the computer readable storage medium would include the following: an electrical connection having one or more wires, a portable computer diskette, 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. In the context of this document, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device.

A computer readable signal medium may include a propagated data signal with computer readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device.

Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

Computer program code for carrying out operations for embodiments of the present invention may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, Smalltalk, 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 computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet service provider).

It is to be noted that the foregoing is only illustrative of the preferred embodiments of the present invention and the technical principles employed. It will be understood by those skilled in the art that the present invention is not limited to the particular embodiments described herein, but is capable of various obvious changes, rearrangements and substitutions as will now become apparent to those skilled in the art without departing from the scope of the invention. Therefore, although the present invention has been described in greater detail by the above embodiments, the present invention is not limited to the above embodiments, and may include other equivalent embodiments without departing from the spirit of the present invention, and the scope of the present invention is determined by the scope of the appended claims.

Claims

1. A method of classifying bronchi in a lung, comprising:

acquiring an initial lung image to be classified for a lung bronchus;

2. The method of claim 1, wherein determining the classification result of the pulmonary bronchi based on the adjacency relationship between the bronchial segments and the segment characteristics of each bronchial segment comprises:

determining adjacency relations among all bronchial segments, and constructing an adjacency matrix corresponding to the pulmonary bronchi based on the adjacency relations;

determining a classification result of each of the bronchial segments based on the adjacency matrix and the segment characteristics of each of the bronchial segments;

determining a classification result of the pulmonary bronchi based on the classification result of each of the bronchial segments.

3. The method of claim 2, wherein determining the classification result for each of the bronchial segments based on the adjacency matrix and the segment characteristics of each of the bronchial segments comprises:

4. The method of claim 2, wherein the segmenting the bronchial tubes of the lungs in the bronchomask image into at least two bronchial segments comprises:

5. The method of claim 4, wherein the segmenting the pulmonary bronchus based on the bronchial midline comprises:

extracting key points of the bronchial midline to obtain at least two target key points of the bronchial midline;

segmenting the pulmonary bronchus based on the at least two target key points.

6. The method of claim 5, wherein the extracting key points of the bronchial centerline to obtain target key points of the bronchial centerline comprises:

7. The method of claim 4, wherein said determining a classification of said pulmonary bronchi based on said classification of each said bronchial segment comprises:

8. A classification device for pulmonary bronchi, comprising:

9. An electronic device, characterized in that the electronic device comprises: one or more processors; storage means for storing one or more programs;

when executed by the one or more processors, cause the one or more processors to implement the method of classifying pulmonary bronchi of any of claims 1-7.

10. A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the method for classifying pulmonary bronchi according to any one of claims 1 to 7.