CN-122025033-A - Airway geometric parameter measurement method and system

CN122025033ACN 122025033 ACN122025033 ACN 122025033ACN-122025033-A

Abstract

The application provides a method and a system for measuring geometrical parameters of an air passage, which relate to the technical field of medical image processing, and the method converts voxel index coordinates into world physical coordinates according to image space meta-information by acquiring air passage foreground voxel data corresponding to chest CT; then backtracking by taking the tail end branch meeting the preset condition as a starting point, calculating the sectional area and the equivalent diameter of the section of each node, fitting the direction vector of the branching node to obtain the branching angle parameter, finally summarizing and outputting a structured file according to algebraic labels and other dimensions, thereby realizing stable measurement and standardized output of the multi-scale geometrical parameters of the air passage and improving the comparability and repeatability of the group scale analysis.

Inventors

GUO WEIQI
XU XINXI
LI PENGHUI
FAN JINBO
ZHAO XIUGUO
SU CHEN
ZHANG YUEHUA

Assignees

军事科学院系统工程研究院卫勤保障技术研究所

Dates

Publication Date: 20260512
Application Date: 20260126

Claims (10)

1. A method for measuring geometrical parameters of an airway, comprising: acquiring airway foreground voxel data corresponding to chest CT, and converting voxel index coordinates of the airway foreground voxel data into world physical coordinates based on image space meta-information; Three-dimensional skeletonization is carried out on the airway foreground voxel data after coordinate conversion to obtain a central line point set, candidate continuous edge relations are established in a preset neighborhood searching range of the central line point set, and a topological tree of the airway central line is generated based on a tree construction strategy with branch penalty items; carrying out branch segmentation and generation coding on the topology tree to obtain a branch segment number and an algebraic label; Tracking the tail end branch of the topological tree to the near end along the central line of the tail end branch, calculating the cross section area of the air passage in a slice where each node is located according to each node on a tracking path, and calculating the cross section equivalent diameter of the node based on the cross section area; In the topology tree, for each bifurcation node, under a world physical coordinate system, based on the central line point set adjacent to the bifurcation node, respectively fitting direction vectors of a parent branch and a child branch to calculate bifurcation angle parameters, wherein the bifurcation angle parameters comprise an included angle between the parent branch and the child branch and an included angle between the child branches; And according to the algebraic labels, the branch segment numbers and the tail end tracking paths, summarizing and counting the airway measurement results, and outputting a structured data file.
2. The method of claim 1, further comprising, after generating the topology tree of the airway centerline: calculating the sub-tree weight of each branch in the topology tree, wherein the sub-tree weight is the sum of the central line physical lengths of all branches on the sub-tree taking the corresponding branch as the root; for each sub-branch under any father node, comparing the sub-tree weight of the target sub-branch with the sum of the sub-tree weights of all sibling sub-branches of the same layer of the target sub-branch; If the subtree weight of the target subtree is lower than the product of the preset pruning threshold and the sum of the subtree weights of all brothers subtrees, judging the corresponding subtree as a pseudo-branch, and deleting the pseudo-branch from the topology tree.
3. The method as recited in claim 1, further comprising: Resampling the airway foreground voxel data to a physical voxel size unified for all subjects when performing population analysis; The world physical coordinates of all centerline points in the topology tree of all subjects, as well as the end world physical coordinates determined by the tracking path, are scaled synchronously according to the resampling scale.
4. The method according to claim 1, wherein the three-dimensionally skeletonizing the coordinate-converted airway foreground voxel data to obtain a centerline point set, establishing candidate continuous edge relations in a preset neighborhood search range of the centerline point set, and generating a topology tree of the airway centerline based on a tree construction strategy with branch penalty terms, comprises: three-dimensional skeletonization is carried out on the airway foreground voxel data after coordinate conversion by applying a three-dimensional morphological refinement algorithm, and a central line point set of single element width is obtained on the premise of keeping an original airway topological structure; For each centerline point in the centerline point set, establishing candidate connection among nodes in a preset neighborhood searching range of the centerline point to construct a weighted undirected graph, wherein the initial weight of each side in the weighted undirected graph is determined based on Euclidean distance between the centerline points at two ends; A topology tree is generated from the weighted undirected graph using a minimum spanning tree algorithm that introduces a branch penalty term.
5. The method of claim 4, wherein generating a topology tree from the weighted undirected graph using a minimum spanning tree algorithm that introduces a branch penalty term, comprises: taking one node in the weighted undirected graph as an initial growth point, and starting to generate a topology tree; when the topology tree is expanded each time, selecting an edge with the smallest comprehensive cost from all candidate edges connecting the accessed node and the non-accessed node, adding the edge into the topology tree, wherein the comprehensive cost of the edge is the sum of a distance cost item and a branch penalty item, and the branch penalty item is activated and counts the comprehensive cost when the degree of departure of the non-accessed node connected by one candidate edge and a father node existing in the topology tree is greater than or equal to 1.
6. The method of claim 1, wherein said encoding the topology tree in branch segments and generations to obtain branch segment numbers and algebraic labels comprises: Traversing the topology tree from a root node of the topology tree; In the traversal process, defining a continuous and non-bifurcated central line path as a branch section, and allocating a unique branch section number to the branch section; Meanwhile, corresponding algebraic labels are distributed to the branch sections according to the number of branch points passing from the root node to the current path.
7. The method of claim 1, further comprising calculating a branch segment equivalent diameter, wherein the branch segment equivalent diameter is calculated for a complete branch segment based on a physical volume of the branch segment and a centerline physical length.
8. The method of claim 1, wherein the airway foreground voxel data is a binary mask, a probability map thresholding result, or a set of voxels equivalent to the binary mask, probability map thresholding result.
9. An airway geometry measurement system, comprising: The acquisition module is used for acquiring airway foreground voxel data corresponding to chest CT and converting voxel index coordinates of the airway foreground voxel data into world physical coordinates based on the image space meta-information; the generation module is used for carrying out three-dimensional skeletonization on the airway foreground voxel data after coordinate conversion to obtain a central line point set, establishing candidate continuous edge relation in a preset neighborhood searching range of the central line point set, and generating a topological tree of the airway central line based on a tree construction strategy with branch punishment items; The segmentation module is used for carrying out branch segmentation and algebraic coding on the topology tree to obtain a branch segment number and algebraic labels; The calculation module is used for taking the tail end branch of the topological tree as a tracking starting point, wherein the tail end branch is a branch section meeting a preset condition, tracking is carried out to the near end along the central line of the tail end branch, the sectional area of the air passage in a slice where each node is positioned is calculated aiming at each node on a tracking path, and the section equivalent diameter of the node is calculated based on the sectional area; the calculation module is further used for respectively fitting direction vectors of a father branch and a son branch based on the central line point set adjacent to the bifurcation node in the topology tree for each bifurcation node under a world physical coordinate system so as to calculate bifurcation angle parameters, wherein the bifurcation angle parameters comprise an included angle between the father branch and the son branch and an included angle between the son branches; and the output module is used for summarizing and counting the airway measurement results according to the algebraic labels, the branch segment numbers and the tail end tracking paths and outputting structured data files.
10. A computer readable storage medium, characterized in that the computer readable storage medium has stored therein a computer program which, when executed by a processor, is capable of implementing the airway geometry parameter measurement method according to any one of claims 1 to 8.

Description

Airway geometric parameter measurement method and system Technical Field The application relates to the technical field of medical image processing, in particular to a method and a system for measuring geometrical parameters of an airway. Background The airway geometric parameter measurement method can convert the airway three-dimensional anatomical structure in the chest CT image into a quantifiable index, is an important tool for respiratory system disease screening, disease assessment and group health research, and has wide application value in clinical diagnosis and medical scientific research. The existing airway geometric parameter measurement technology directly extracts the central line of the airway and calculates basic parameters such as length, diameter and the like after voxel data corresponding to the airway are acquired through image segmentation, a coordinate system is not required to be uniformly calibrated, the central line structure is not optimized, and a measurement result is only output in a scattered data form. Meanwhile, the central line extraction process is susceptible to noise to generate pseudo branches, so that the measurement of key parameters such as diameter, bifurcation angle and the like is inaccurate, and the data of different individuals or researches are difficult to directly compare due to the fact that a unified branch classification standard and a structured output format are not available. Therefore, the technical problems of difficult comparison caused by insufficient measurement accuracy of the geometric parameters of the air passages and lack of unified specification of data exist in the prior art. Disclosure of Invention The application aims to provide a method and a system for measuring airway geometric parameters, which are used for solving the problem of difficulty in comparison caused by insufficient accuracy and lack of unified and standard data of airway geometric parameters in the prior art. To solve the above technical problem, in a first aspect, the present application provides a method for measuring geometric parameters of an airway, including: acquiring airway foreground voxel data corresponding to chest CT, and converting voxel index coordinates of the airway foreground voxel data into world physical coordinates based on image space meta-information; Three-dimensional skeletonization is carried out on the airway foreground voxel data after coordinate conversion to obtain a central line point set, candidate continuous edge relations are established in a preset neighborhood searching range of the central line point set, and a topological tree of the airway central line is generated based on a tree construction strategy with branch penalty items; carrying out branch segmentation and generation coding on the topology tree to obtain a branch segment number and an algebraic label; Tracking the tail end branch of the topological tree to the near end along the central line of the tail end branch, calculating the cross section area of the air passage in a slice where each node is located according to each node on a tracking path, and calculating the cross section equivalent diameter of the node based on the cross section area; In the topology tree, for each bifurcation node, under a world physical coordinate system, based on the central line point set adjacent to the bifurcation node, respectively fitting direction vectors of a parent branch and a child branch to calculate bifurcation angle parameters, wherein the bifurcation angle parameters comprise an included angle between the parent branch and the child branch and an included angle between the child branches; And according to the algebraic labels, the branch segment numbers and the tail end tracking paths, summarizing and counting the airway measurement results, and outputting a structured data file. Optionally, after generating the topology tree of the airway centerline, further comprising: calculating the sub-tree weight of each branch in the topology tree, wherein the sub-tree weight is the sum of the central line physical lengths of all branches on the sub-tree taking the corresponding branch as the root; for each sub-branch under any father node, comparing the sub-tree weight of the target sub-branch with the sum of the sub-tree weights of all sibling sub-branches of the same layer of the target sub-branch; If the subtree weight of the target subtree is lower than the product of the preset pruning threshold and the sum of the subtree weights of all brothers subtrees, judging the corresponding subtree as a pseudo-branch, and deleting the pseudo-branch from the topology tree. Optionally, the method further comprises: Resampling the airway foreground voxel data to a physical voxel size unified for all subjects when performing population analysis; The world physical coordinates of all centerline points in the topology tree of all subjects, as well as the end world physical coordinates determined by the tr