A three-dimensional reconstruction method for chest images of patients with early-stage lung cancer

Abstract

The present invention relates to the field of image data processing technology, and more specifically to a method for three-dimensional reconstruction of chest images of patients with early-stage lung cancer. The method comprises: obtaining a lung region from each lung slice image in a sequence of lung slice images of the patient with early-stage lung cancer; segmenting a plurality of low-density tissue connected domains from the lung region; determining the likelihood of a pathological condition in the low-density tissue connected domains based on the area differences of the low-density tissue connected domains between consecutive adjacent lung slice images; thereby distinguishing the low-density tissue connected domains into diseased tissue connected domains and airway tissue connected domains, assigning different labels, and converting the labeled lung slice image sequence into a three-dimensional model. By accurately distinguishing diseased tissue from airway tissue, the present invention improves the ability of the three-dimensional reconstruction model to express early-stage diseased tissue.

Description

Technical Field

[0001] The present invention relates to the technical field of image data processing, and in particular to a method for three-dimensional reconstruction of chest images of patients with early-stage lung cancer. Background Art

[0002] Three-dimensional reconstruction of chest CT images of lung cancer patients is an important medical imaging technology. By processing and analyzing the two-dimensional slice images obtained by CT scans to generate three-dimensional stereo images, it can effectively help doctors more comprehensively assess the patient's lung condition, including the size and location of the tumor and its relationship to surrounding tissues, providing key basis for subsequent surgical treatment plans.

[0003] Existing Problem: 3D reconstruction based on CT images requires accurate threshold segmentation of discrete 2D slice images. After accurately segmenting the airway tissue and other lesions, the images are converted into pixels in 3D coordinates for 3D reconstruction. However, lesions in early-stage lung cancer patients are irregular and small in size, and appear similar to other normal airway tissue on CT images. Therefore, effective differentiation of these lesions can be difficult, resulting in low accuracy in the final 3D reconstructed model in representing the lesions. Summary of the Invention

[0004] The present invention provides a three-dimensional reconstruction method for chest images of patients with early lung cancer to solve the existing problems.

[0005] The present invention provides a method for three-dimensional reconstruction of chest images of patients with early-stage lung cancer using the following technical solutions:

[0006] One embodiment of the present invention provides a method for three-dimensional reconstruction of chest images of patients with early-stage lung cancer, the method comprising the following steps:

[0007] Acquire a sequence of lung slice images of a patient with early-stage lung cancer; and segment the lung region in each lung slice image using a segmentation neural network in the sequence of lung slice images;

[0008] Segmenting a plurality of low-density tissue connected domains from the lung region in each lung slice image according to grayscale value differences between pixels in the lung region in each lung slice image;

[0009] determining the lesion possibility of each low-density tissue connected domain in each lung slice image based on a difference in the area of the low-density tissue connected domain between consecutive adjacent lung slice images in the lung slice image sequence;

[0010] According to the pathological possibility of each low-density tissue connected domain in each lung slice image, the low-density tissue connected domain is divided into a diseased tissue connected domain and an airway tissue connected domain; different labels are assigned to the diseased tissue connected domain and the airway tissue connected domain in each lung slice image in the lung slice image sequence to obtain a labeled lung slice image sequence; and the labeled lung slice image sequence is converted into a three-dimensional model.

[0011] Furthermore, the segmenting of a plurality of low-density tissue connected domains from the lung region in each lung slice image includes the following specific steps:

[0012] In the lung slice image sequence, a lung region in any lung slice image is recorded as a target lung region;

[0013] Segmenting a plurality of initial airway tissue connected domains from the target lung region;

[0014] The boundary pixel points on each initial airway tissue connected domain are recorded as seed points;

[0015] Preset several attenuation directions, according to The initial airway tissue connectivity domain along the The gray value difference of the pixel points in the preset attenuation direction is used to determine the The initial airway tissue connectivity domain is The attenuation degree in a preset attenuation direction;

[0016] According to The attenuation degree of the initial airway tissue connectivity domain in all preset attenuation directions is determined by The similarity threshold of each seed point on the initial airway tissue connectivity domain;

[0017] In the target lung area, a region growing algorithm is used based on each seed point and the similarity criterion of each seed point to start region growing from each seed point to obtain several connected domains as low-density tissue connected domains; the similarity criterion of each seed point is: when the absolute value of the difference between the grayscale value of each seed point and each pixel point adjacent to each seed point is less than the similarity threshold of each seed point, each adjacent pixel point is merged into the region where each seed point is located.

[0018] Furthermore, the segmenting of a plurality of initial airway tissue connected domains from the target lung region includes the following specific steps:

[0019] The Otsu algorithm is used to obtain the optimal segmentation threshold of the target lung area. In the target lung area, pixels with grayscale values greater than the optimal segmentation threshold are recorded as target pixels, and the connected domain formed by adjacent target pixels is recorded as the initial airway tissue connected domain.

[0020] Furthermore, the determination The initial airway tissue connectivity domain is The attenuation degree in a preset attenuation direction includes the following specific steps:

[0021] The first The vertical direction of the preset attenuation direction is used as the reference direction;

[0022] First A straight line in the preset attenuation direction, The initial airway tissue connection domain is translated along the reference direction to traverse the entire Initial airway tissue connectivity domains are obtained, and several non-overlapping straight lines are obtained as target straight lines;

[0023] In the In the initial airway tissue connectivity domain, all pixels on each target line are obtained. A sequence consisting of a preset attenuation direction as the positive direction is used as the pixel sequence of each target line;

[0024] In the pixel sequence of each target line, calculate the The gray value of the pixel is subtracted from the The difference in the grayscale values of pixels is taken as the average of the differences in the grayscale values of all adjacent pixels as the attenuation of each target line;

[0025] The mean of the attenuation of all target lines is taken as the The initial airway tissue connectivity domain is The attenuation degree in each preset attenuation direction.

[0026] Furthermore, the determination The similarity threshold of each seed point on the initial airway tissue connectivity domain includes the following specific steps:

[0027] The first The maximum attenuation degree of the initial airway tissue connectivity domain in all preset attenuation directions is recorded as The similarity threshold of each seed point on the initial airway tissue connectivity domain.

[0028] Furthermore, the step of determining the lesion possibility of each low-density tissue connected domain in each lung slice image includes the following specific steps:

[0029] In the lung slice image sequence, according to A plurality of adjacent lung slice images before and after the first lung slice image are obtained. a reference lung slice image sequence segment of a lung slice image;

[0030] The first Any low-density tissue connected domain in the lung slice image is recorded as the target low-density tissue connected domain;

[0031] acquiring an area sequence segment according to the area of the low-density tissue connected domain matched to the target low-density tissue connected domain in the reference lung slice image sequence segment;

[0032] According to the difference between the areas in the area sequence segments, the first The lesion probability of the target low-density tissue connected area in the lung slice image.

[0033] Further, the obtaining of the The specific steps of generating a reference lung slice image sequence segment of a lung slice image are as follows:

[0034] Preset quantity threshold In the lung slice image sequence, lung slice images to The sequence segment consisting of lung slice images is A reference lung slice image sequence segment of a lung slice image.

[0035] Furthermore, the specific steps of obtaining the area sequence segments are as follows:

[0036] In the said In a reference lung slice image sequence segment of a plurality of lung slice images, an optical flow method is used to obtain a matching target low-density tissue connected domain of the target low-density tissue connected domain in each reference lung slice image to obtain a matching target low-density tissue connected domain sequence segment;

[0037] In the matching target low-density tissue connected domain sequence segments, the number of pixels in each matching target low-density tissue connected domain is counted as the area of each matching target low-density tissue connected domain, and an area sequence segment is obtained.

[0038] Furthermore, the determination The specific steps of determining the pathological possibility of the target low-density tissue connected area in the lung slice image are as follows:

[0039] In the area sequence segment, the variance of all areas is calculated, and then the absolute value of the difference between adjacent areas is calculated. The normalized value of the product of the mean of the absolute value of the difference between all adjacent areas and the variance is recorded as the first The lesion probability of the target low-density tissue connected area in the lung slice image.

[0040] Furthermore, the step of dividing the low-density tissue connected domain into the diseased tissue connected domain and the airway tissue connected domain comprises the following specific steps:

[0041] The low-density tissue connected area with a lesion possibility greater than the preset judgment threshold is recorded as the lesion tissue connected area;

[0042] The low-density tissue connectivity domain with a lesion possibility less than or equal to the preset judgment threshold is recorded as the airway tissue connectivity domain.

[0043] The beneficial effects of the technical solution of the present invention are:

[0044] In an embodiment of the present invention, the lung area in each lung slice image in a sequence of lung slice images of a patient with early lung cancer is obtained, and a number of low-density tissue connected domains are segmented from the lung area, thereby ensuring the accuracy of subsequent differentiation of diseased tissues by accurately segmenting the low-density tissue connected domains containing diseased tissues and airway tissues. According to the difference in the area of the low-density tissue connected domains between consecutive adjacent lung slice images, the possibility of the disease in the low-density tissue connected domain is determined, thereby dividing the low-density tissue connected domain into diseased tissue connected domains and airway tissue connected domains, and assigning different labels, thereby further ensuring the accuracy of diseased tissue segmentation based on the analysis between consecutive lung slice images. The labeled lung slice image sequence is converted into a three-dimensional model. The present invention can effectively help determine the location of diseased tissue in the three-dimensional reconstruction process by accurately distinguishing diseased tissue from airway tissue, thereby improving the ability of the three-dimensional reconstruction model to express early diseased tissue. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0046] Figure 1 This is a flowchart of the steps of a method for three-dimensional reconstruction of chest images of patients with early lung cancer according to the present invention;

[0047] Figure 2 This is a schematic diagram of chest CT imaging;

[0048] Figure 3 This is a CT slice image of a lung nodule in the early stage of lung cancer;

[0049] Figure 4 This is a slice image that includes only the lungs;

[0050] Figure 5 This is the image of the lung area after segmentation using the Otsu algorithm;

[0051] Figure 6 Schematic diagram of the division of the connected domain pixel sequence with the upper left as the attenuation direction;

[0052] Figure 7 This is a three-dimensional schematic diagram of a CT slice of lung airway tissue. DETAILED DESCRIPTION

[0053] To further illustrate the technical means and effectiveness of the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, describes in detail a method for 3D reconstruction of chest images for patients with early-stage lung cancer, including its specific implementation, structure, features, and effectiveness. In the following description, different references to "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.

[0054] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0055] The following describes in detail a method for three-dimensional reconstruction of chest images of patients with early-stage lung cancer provided by the present invention with reference to the accompanying drawings.

[0056] See also Figure 1 , which shows a flowchart of a method for three-dimensional reconstruction of chest images of patients with early lung cancer provided by one embodiment of the present invention, the method comprising the following steps:

[0057] Step S001: Acquire a sequence of lung slice images of a patient with early-stage lung cancer; in the sequence of lung slice images, use a segmentation neural network to segment the lung region in each lung slice image.

[0058] Using a 128-slice spiral CT device, chest images of patients with early-stage lung cancer are collected to obtain a sequence of lung slice images of patients with early-stage lung cancer.

[0059] What needs to be explained is: Schematic diagram of chest CT shooting, such as Figure 2 As shown, Figure 2The 128-slice CT scanner is a technical parameter indicating that the scanner can simultaneously acquire image data from 128 different slices in a single rotation. In this embodiment, the slice thickness of the lung slice images is set to 1 mm, and the lung slice image sequence contains 200 grayscale images. This example is used for the purpose of this description.

[0060] In most cases, early-stage lung cancer presents as pulmonary nodules, which can take on a variety of shapes. For example, benign tumors are often ellipsoidal, while malignant tumors are often spiculated and lobed. Cancerous tissue is generally smaller and less dense. In lung slice images, normal airway tissue, due to the presence of air, also has a lower density. Consequently, the grayscale of cancerous and normal airway tissue appears similar in lung slice images, making it difficult to distinguish. Consequently, the cancerous tissue can be easily overlooked during 3D reconstruction.

[0061] However, because normal airway tissue has a tree-like bifurcated structure, it may appear discontinuous in a lung slice image, while multiple adjacent lung slice images can form continuous airway tissue. Since the diameter width gradually decreases from the main airway to the branch airway, early cancerous lung nodules are usually irregular in structure. Although they may appear in various forms such as burrs or lobes, their continuity in continuous lung slice images is much lower than that of normal tissue, and their width is irregular. CT slice images of early lung cancer nodules, such as Figure 3 shown.

[0062] Therefore, in this embodiment, the lung slice image is first preliminarily segmented to determine the outer contour of the lung and the low-density tissue inside. Secondly, the possibility that the low-density tissue is normal airway tissue is calculated. Finally, a segmentation threshold is set to distinguish normal airway tissue from cancerous tissue.

[0063] Because the captured CT image contains all objects in the scanned area, including bed boards, clothing, and other human tissue, the lung region needs to be extracted and then initially segmented to obtain the low-density tissue within the lung region. Because the low-density tissue inside the lung (airway tissue or cancerous tissue) interferes with the clothing, bed boards, and other objects outside the lung, direct image segmentation is ineffective. However, the lung region has a clear and continuous grayscale boundary, making it easy to identify. Therefore, the lung region is identified first, followed by the low-density tissue within the lung.

[0064] In a sequence of lung slice images, taking any lung slice image as an example, an embodiment of the present invention uses a segmentation neural network to identify and segment the lung region in the lung slice image.

[0065] The relevant content of the segmentation neural network is as follows:

[0066] The segmentation neural network used in this example is the Mask R-CNN neural network, and the dataset used is a lung slice image dataset. Mask R-CNN is a well-known technique, and the specific method is not described here. Mask R-CNN stands for "Mask Region-based Convolutional Neural Network" in Chinese and "Mask Region-based Convolutional Neural Network" in English.

[0067] The pixels that need to be segmented are divided into two categories, that is, the labeling process of the training set is: single-channel semantic label, the corresponding position pixel belongs to the background area is labeled as 0, and belongs to the lung area is labeled as 1.

[0068] The task of the network is classification, so the loss function used is the cross entropy loss function.

[0069] The lung region in each lung slice image is obtained by segmentation neural network. This process is well known and the specific method will not be introduced here. Figure 4 shown.

[0070] Step S002: Segmenting a plurality of low-density tissue connected domains from the lung region in each lung slice image according to the grayscale value differences between the pixels in the lung region in each lung slice image.

[0071] Because airway tissue or cancerous tissue radiates outward, there is divergent tissue in lung slice images, where the diameter narrows from the main end to the terminal end, and the image grayscale gradually decreases, which is the cross-section of the airway. There is also tissue that penetrates the slice, which is the longitudinal section of the airway. These tissues may appear as isolated grayscale points within a slice plane, or even tissue that combines the two characteristics. Therefore, the airway tissue types in lung slice images can be divided into divergent within the section, penetrating the section, and the coexistence of penetrating and diverging within the section. When segmenting lung slice images, airway tissue and cancerous tissue that diverge within the section are easily unable to be completely segmented, and only partial tissue with a grayscale similar to the penetrating type can be segmented. Therefore, it is necessary to first determine the complete airway tissue.

[0072] Preferably, in one embodiment of the present invention, the method for obtaining the low-density tissue connected area includes:

[0073] In the lung slice image sequence, a lung region in any lung slice image is recorded as a target lung region.

[0074] The Otsu algorithm was used to obtain the optimal segmentation threshold of the target lung area. In the target lung area, pixels with grayscale values greater than the optimal segmentation threshold were recorded as target pixels, the connected domain formed by adjacent target pixels was recorded as the initial airway tissue connected domain, and the boundary pixels on each initial airway tissue connected domain were recorded as seed points.

[0075] It should be noted that in this embodiment, the region growing algorithm is used to obtain the low-density tissue region in the lung region, which includes airway tissue and cancerous tissue. The region growing algorithm and the Otsu algorithm are both well-known technologies, and the specific methods are not introduced here. The image of the lung region after segmentation by the Otsu algorithm is as follows: Figure 5 shown. Figure 5 The white connected domain within the mid-lung region represents areas with more pronounced airway tissue, known as the initial airway tissue connected domain. Therefore, the boundary pixels within this initial airway tissue connected domain are used as seed points during region growing. Since airway tissue or spicate-shaped cancerous tissue on a cross-section typically appears linear, with grayscale attenuation from one side to the other, and the diameter of airway tissue can be considered uniformly narrowed within a small range, resulting in similar grayscale attenuation, we can analyze the attenuation in each direction within the initial airway tissue connected domain, select the maximum attenuation, and determine the similarity criterion during region growing.

[0076] Several attenuation directions are preset. In this embodiment, eight neighborhood directions, namely, up, down, left, right, upper left, lower left, upper right, and lower right, are used as preset attenuation directions, and this is used as an example for description.

[0077] In the target lung area, For example, the initial airway tissue connectivity domain For example, take the preset attenuation direction The vertical direction of the preset attenuation direction is used as the reference direction. A straight line in the preset attenuation direction, In the initial airway tissue connected domain, translate along the reference direction and traverse the entire Initial airway tissue connectivity domain, obtain several non-overlapping straight lines as target straight lines. In the initial airway tissue connectivity domain, all pixels on each target line are obtained. A sequence consisting of the preset attenuation directions as the positive direction is used as the pixel sequence of each target line.

[0078] It should be noted that if the length of a pixel sequence is 1, it will be discarded and no subsequent analysis will be performed. The diagram of the connected domain pixel sequence division with the upper left as the attenuation direction is as follows: Figure 6 shown. Figure 6The medium gray squares 1, 2, 3, 4, 5, 6, and 7 are connected domain pixels, that is, the gray area is the initial airway tissue connected domain. Taking the upper left direction pointed by the dotted arrow as the attenuation direction, three pixel sequences {2, 1}, {5, 4, 3}, and {7, 6} can be obtained.

[0079] In the pixel sequence of each target line, calculate the The gray value of the pixel is subtracted from the The difference of the gray value of each pixel point is taken as the average of the gray value differences of all adjacent pixels as the attenuation of each target line, and the average of the attenuation of all target lines is taken as the average of the attenuation of the first target line. The initial airway tissue connectivity domain is The attenuation degree in each preset attenuation direction.

[0080] According to the above method, obtain the The attenuation degree of the initial airway tissue connectivity domain in each preset attenuation direction.

[0081] The first The maximum attenuation degree of the initial airway tissue connectivity domain in all preset attenuation directions is recorded as The similarity threshold of each seed point on the initial airway tissue connectivity domain.

[0082] According to the above method, the similarity threshold of each seed point in the target lung region is obtained.

[0083] Within the target lung region, a region growing algorithm is used based on each seed point and its similarity criterion. Several connected domains are obtained as low-density tissue connected domains, starting from each seed point. The similarity criterion for each seed point is: when the absolute value of the grayscale difference between each seed point and each adjacent pixel is less than the similarity threshold for each seed point, each adjacent pixel is merged into the region where the seed point resides.

[0084] It should be noted that the region growing algorithm is a well-known technique. The selection of seed points and the definition of similarity criteria for each seed point are the key operations of the region growing algorithm. During seed point growth, if the difference between the growing pixel and the seed pixel is less than the maximum attenuation, it is considered to be caused by changes in the airway or cancerous tissue, and growth continues. Otherwise, it is considered to be the boundary between airway tissue and other lung tissue, and growth stops.

[0085] According to the above method, a plurality of low-density tissue connected domains in the lung region of each lung slice image in the lung slice image sequence are obtained.

[0086] At this point, the low-density tissue of the airway or lesion in the lung area in each lung slice image is completely segmented.

[0087] Step S003: determining the lesion possibility of each low-density tissue connected domain in each lung slice image based on the difference in area of the low-density tissue connected domain between consecutive adjacent lung slice images in the lung slice image sequence.

[0088] In the tissue with lower density, there are normal airway tissue and suspected cancerous tissue. However, the normal airway tissue presents a bifurcated tree crown structure in the three-dimensional structure. On the CT slices of adjacent intervals, the airway tissue at similar positions can form a continuous tissue, and the tissue width varies from small to large. Figure 7 shown. Figure 7 The continuous diamond boxes in the middle are continuous lung CT slices. The solid lines and dotted lines both represent airway tissue. The solid line represents the distribution of the airway in the direction of the slice, and the dotted line represents the airway tissue in the same slice direction.

[0089] Correspondingly, the continuity of diseased tissue in consecutive slices is not obvious, and the width change continuity is poor. Therefore, if the current low-density tissue has poor continuity in adjacent lung slice images and the width change is not continuous in a single direction, it is highly likely to be diseased tissue.

[0090] In a sequence of lung slice images, the size of the lung region outline varies little across locally consecutive images. The width of the tissue duct is also reflected in the number of pixels within the tissue region; a greater number of pixels indicates a wider tissue duct. Furthermore, in the direction of low-density tissue running through the slice, locally consecutive slices often represent the same tissue at the same location. Therefore, tissue with a regular change in the number of pixels within a connected domain at the same location on consecutive slices is more likely to be normal and less likely to be diseased.

[0091] Preferably, in one embodiment of the present invention, the method for obtaining the lesion possibility of each low-density tissue connectivity domain includes:

[0092] Preset quantity threshold The value is 3, and this is used as an example for description.

[0093] In the lung slice image sequence, Take the lung slice image as an example. lung slice images to The sequence segment consisting of lung slice images is A reference lung slice image sequence segment of a lung slice image.

[0094] What needs to be explained is: The lung slice image is the first lung slice image in the lung slice image sequence. Zhang or last Zhang Shi, in the lung slice images to The existing lung slice images are selected between the lung slice images to form a reference lung slice image sequence segment.

[0095] The first Any low-density tissue connected domain in a lung slice image is recorded as a target low-density tissue connected domain.

[0096] In the In the reference lung slice image sequence of the first lung slice image, the optical flow method is used to obtain the matching target low-density tissue connected domain in each reference lung slice image, and the first Matching target low-density tissue connected domain sequence segments of a lung slice image.

[0097] In the In the sequence of matching target low-density tissue connected domains of the lung slice image, the number of pixels in each matching target low-density tissue connected domain is counted as the area of each matching target low-density tissue connected domain, and the first The area sequence segments corresponding to the target low-density tissue connected domain in the lung slice image.

[0098] It should be noted that the optical flow method is a well-known technique, and the specific method will not be described here. This algorithm can be used to track changes in lung tissue in consecutive slice images. When a target low-density tissue connected domain does not have a matching target low-density tissue connected domain in a reference lung slice image, the number of pixels in the matching target low-density tissue connected domain in the reference lung slice image is set to 0.

[0099] In the In the area sequence segment corresponding to the target low-density tissue connected domain in the lung slice image, the variance of all areas is calculated, and then the absolute value of the difference between adjacent areas is calculated. The normalized value of the product of the mean of the absolute value of the difference between all adjacent areas and the variance is recorded as the first The lesion probability of the target low-density tissue connected area in the lung slice image.

[0100] It should be noted that: the normalized value of the product is used in this embodiment The linear normalization function is used to normalize the product to between 0 and 1. When the above variance is smaller and the absolute value of the difference between adjacent areas is smaller, it means that the first The more uniformly the area of the target low-density tissue connected domain in the lung slice image changes in the continuous slice images, the less likely it is to be a lesion.

[0101] According to the above method, the lesion possibility of each low-density tissue connected domain in each lung slice image in the lung slice image sequence is obtained.

[0102] Step S004: Based on the pathological possibility of each low-density tissue connected domain in each lung slice image, the low-density tissue connected domain is divided into a diseased tissue connected domain and an airway tissue connected domain; different labels are assigned to the diseased tissue connected domain and the airway tissue connected domain in each lung slice image in the lung slice image sequence to obtain a labeled lung slice image sequence; and the labeled lung slice image sequence is converted into a three-dimensional model.

[0103] Preferably, in one embodiment of the present invention, the method for obtaining the three-dimensional model includes:

[0104] The preset judgment threshold is 0.8, and this is used as an example for description.

[0105] The low-density tissue connected area with a lesion possibility greater than the preset judgment threshold is recorded as the lesion tissue connected area.

[0106] The low-density tissue connectivity domain with a lesion possibility less than or equal to the preset judgment threshold is recorded as the airway tissue connectivity domain.

[0107] In the lung slice image sequence, different labels are assigned to the diseased tissue connected domain and the airway tissue connected domain in each lung slice image, and the labeled lung slice image sequence is obtained. The labeled lung slice image sequence is converted into a three-dimensional model using a three-dimensional reconstruction algorithm.

[0108] It should be noted that the three-dimensional reconstruction algorithm is a well-known technology and the specific method will not be introduced here. In this embodiment, the diseased tissue connected area in each lung slice image is marked in red, and the airway tissue connected area is marked in yellow. This is used as an example for description.

[0109] So far, the present invention is completed.

[0110] In summary, in an embodiment of the present invention, a lung region is obtained from each lung slice image in a sequence of lung slice images of a patient with early-stage lung cancer, and a number of low-density tissue connected domains are segmented from the lung region. The possibility of a lesion in the low-density tissue connected domain is determined based on the difference in the area of the low-density tissue connected domains between consecutive adjacent lung slice images. The low-density tissue connected domains are then divided into lesion tissue connected domains and airway tissue connected domains, and different labels are assigned to the low-density tissue connected domains. The labeled lung slice image sequence is then converted into a three-dimensional model. By accurately distinguishing lesion tissue from airway tissue, the present invention improves the ability of the three-dimensional reconstructed model to express early-stage lesion tissue.

[0111] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for three-dimensional reconstruction of chest images of patients with early-stage lung cancer, characterized in that: The method comprises the following steps: Acquire a sequence of lung slice images of a patient with early-stage lung cancer; and segment the lung region in each lung slice image using a segmentation neural network in the sequence of lung slice images; Segmenting a plurality of low-density tissue connected domains from the lung region in each lung slice image according to grayscale value differences between pixels in the lung region in each lung slice image; determining the lesion possibility of each low-density tissue connected domain in each lung slice image based on the difference in area of the low-density tissue connected domains between adjacent lung slice images in the lung slice image sequence; According to the pathological possibility of each low-density tissue connected domain in each lung slice image, the low-density tissue connected domain is divided into a diseased tissue connected domain and an airway tissue connected domain; different labels are assigned to the diseased tissue connected domain and the airway tissue connected domain in each lung slice image in the lung slice image sequence to obtain a labeled lung slice image sequence; and the labeled lung slice image sequence is converted into a three-dimensional model.

2. The method for three-dimensional reconstruction of chest images of patients with early-stage lung cancer according to claim 1, characterized in that: The specific steps of segmenting a plurality of low-density tissue connected domains from the lung region in each lung slice image are as follows: In the lung slice image sequence, a lung region in any lung slice image is recorded as a target lung region; Segmenting a plurality of initial airway tissue connected domains from the target lung region; The boundary pixel points on each initial airway tissue connected domain are recorded as seed points; Preset several attenuation directions, according to The initial airway tissue connectivity domain along the The gray value difference of the pixel points in the preset attenuation direction is used to determine the The initial airway tissue connectivity domain is The attenuation degree in a preset attenuation direction; According to The attenuation degree of the initial airway tissue connectivity domain in all preset attenuation directions is determined by The similarity threshold of each seed point on the initial airway tissue connectivity domain; In the target lung area, a region growing algorithm is used based on each seed point and the similarity criterion of each seed point to start region growing from each seed point to obtain several connected domains as low-density tissue connected domains; the similarity criterion of each seed point is: when the absolute value of the difference between the grayscale value of each seed point and each pixel point adjacent to each seed point is less than the similarity threshold of each seed point, each adjacent pixel point is merged into the region where each seed point is located.

3. The method for three-dimensional reconstruction of chest images of patients with early-stage lung cancer according to claim 2, characterized in that: The step of segmenting a plurality of initial airway tissue connected regions from the target lung region includes the following specific steps: The Otsu algorithm is used to obtain the optimal segmentation threshold of the target lung area. In the target lung area, pixels with grayscale values greater than the optimal segmentation threshold are recorded as target pixels, and the connected domain formed by adjacent target pixels is recorded as the initial airway tissue connected domain.

4. The method for three-dimensional reconstruction of chest images of patients with early-stage lung cancer according to claim 2, characterized in that: The determination of The initial airway tissue connectivity domain is The attenuation degree in a preset attenuation direction includes the following specific steps: The first The vertical direction of the preset attenuation direction is used as the reference direction; First A straight line in the preset attenuation direction, The initial airway tissue connection domain is translated along the reference direction to traverse the entire Initial airway tissue connectivity domains are obtained, and several non-overlapping straight lines are obtained as target straight lines; In the In the initial airway tissue connectivity domain, all pixels on each target line are obtained. A sequence consisting of a preset attenuation direction as the positive direction is used as the pixel sequence of each target line; In the pixel sequence of each target line, calculate the The gray value of the pixel is subtracted from the The difference in the grayscale values of pixels is taken as the average of the differences in the grayscale values of all adjacent pixels as the attenuation of each target line; The mean of the attenuation of all target lines is taken as the The initial airway tissue connectivity domain is The attenuation degree in each preset attenuation direction.

5. The method for three-dimensional reconstruction of chest images of patients with early-stage lung cancer according to claim 2, characterized in that: The determination of The similarity threshold of each seed point on the initial airway tissue connectivity domain includes the following specific steps: The first The maximum attenuation degree of the initial airway tissue connectivity domain in all preset attenuation directions is recorded as The similarity threshold of each seed point on the initial airway tissue connectivity domain.

6. The method for three-dimensional reconstruction of chest images of patients with early-stage lung cancer according to claim 1, characterized in that: The specific steps of determining the lesion possibility of each low-density tissue connected domain in each lung slice image are as follows: In the lung slice image sequence, according to A plurality of adjacent lung slice images before and after the first lung slice image are obtained. a reference lung slice image sequence segment of a lung slice image; The first Any low-density tissue connected domain in the lung slice image is recorded as the target low-density tissue connected domain; acquiring an area sequence segment according to the area of the low-density tissue connected domain matched to the target low-density tissue connected domain in the reference lung slice image sequence segment; According to the difference between the areas in the area sequence segments, the first The lesion probability of the target low-density tissue connected area in the lung slice image.

7. The method for three-dimensional reconstruction of chest images of patients with early-stage lung cancer according to claim 6, characterized in that: The acquisition of the The specific steps of generating a reference lung slice image sequence segment of a lung slice image are as follows: Preset quantity threshold In the lung slice image sequence, lung slice images to The sequence segment consisting of lung slice images is A reference lung slice image sequence segment of a lung slice image.

8. The method for three-dimensional reconstruction of chest images of patients with early-stage lung cancer according to claim 6, characterized in that: The specific steps of obtaining the area sequence segment are as follows: In the said In a reference lung slice image sequence segment of a plurality of lung slice images, an optical flow method is used to obtain a matching target low-density tissue connected domain of the target low-density tissue connected domain in each reference lung slice image to obtain a matching target low-density tissue connected domain sequence segment; In the matching target low-density tissue connected domain sequence segments, the number of pixels in each matching target low-density tissue connected domain is counted as the area of each matching target low-density tissue connected domain, and an area sequence segment is obtained.

9. The method for three-dimensional reconstruction of chest images of patients with early-stage lung cancer according to claim 6, characterized in that: The determination of The specific steps involved in determining the pathological possibility of the target low-density tissue connected region in the lung slice image are as follows: In the area sequence segment, the variance of all areas is calculated, and then the absolute value of the difference between adjacent areas is calculated. The normalized value of the product of the mean of the absolute value of the difference between all adjacent areas and the variance is recorded as the first The lesion probability of the target low-density tissue connected area in the lung slice image.

10. The method for three-dimensional reconstruction of chest images of patients with early-stage lung cancer according to claim 1, characterized in that: The method of dividing the low-density tissue connected area into the lesion tissue connected area and the airway tissue connected area includes the following specific steps: The low-density tissue connected area with a lesion possibility greater than the preset judgment threshold is recorded as the lesion tissue connected area; The low-density tissue connectivity domain with a lesion possibility less than or equal to the preset judgment threshold is recorded as the airway tissue connectivity domain.

Images