Classification method and analysis device for cerebrovascular branch in cerebrovascular image
A method of classifying a cerebrovascular branch in cerebrovascular image is disclosed, the method including receiving, by an analysis device, an cerebrovascular image of a subject; extracting, by the analysis device, a plurality of vascular unit structures from the cerebrovascular image based on geometric features of a 3D model; extracting, by the analysis device, feature values for each of the plurality of vascular unit structures; inputting, by the analysis device, the feature values of each of the plurality of vascular unit structures into a first learning model trained in advance, and classifying chunks to which each of the plurality of vascular unit structures belongs; and inputting, by the analysis device, the feature values of each of the vascular unit structures belonging to same chunk into a second learning model trained in advance, and classifying a vessel branch composed of the vascular unit structures belonging to the same chunk.
1 . A method of classifying a cerebrovascular branch in cerebrovascular image, the method comprising:
receiving, by an analysis device, a cerebrovascular image of a subject;
extracting, by the analysis device, a plurality of vascular unit structures from the cerebrovascular image based on geometric features of a three-dimensional (3D) model;
extracting, by the analysis device, feature values for each of the plurality of vascular unit structures;
inputting, by the analysis device, the feature values of each of the plurality of vascular unit structures into a first learning model trained in advance, and classifying chunks to which each of the plurality of vascular unit structures belongs; and
inputting, by the analysis device, the feature values of each of the vascular unit structures belonging to same chunk into a second learning model trained in advance, and classifying a vessel branch composed of the vascular unit structures belonging to the same chunk, and
wherein each vascular unit structure is a spot, and the spot is a cell having a regular spacing from the artery center line extracted from the cerebrovascular image.
2 . The method of claim 1 , wherein the chunk is, as a higher structure of a vessel branch including at least one vessel branch type, a vascular unit structure segmented into different types according to at least one criterion of (i) symmetry; (ii) anterior or posterior; (iii) basal or pial; and (iv) a group including criteria of middle cerebral arteries (MCA), anterior cerebral arteries (ACA) and posterior cerebral arteries (PCA).
3 . The method of claim 1 , wherein the feature values include cerebral vessel cross-sectional area, maximum inscribed sphere radius, minimum diameter, maximum diameter, maximum-minimum radius ratio, surface circumference, distortion, curvature, and lumen roundness.
4 . The method of claim 3 , wherein the feature values further include a brightness value of the vascular unit structure.
5 . The method of claim 1 , wherein the classifying of the chunks includes:
performing, by the analysis device, primary chunk classification for each of the plurality of vascular unit structures using the first learning model; and
performing, by the analysis device, secondary chunk classification for the vascular unit structures belonging to same segment, in a majority voting manner based on results of the primary chunk classification of the vascular unit structures belonging to the same segment among the plurality of vascular unit structures,
wherein the segment is composed of vascular unit structures belonging to a region segmented by a branch point in a vascular structure.
6 . The method of claim 1 , wherein the classifying of the vessel branch includes:
performing, by the analysis device, primary vessel branch classification for each of the plurality of vascular unit structures belonging to same chunk using the second learning model; and
performing, by the analysis device, secondary vessel branch classification for the vascular unit structures belonging to same chunk, in a majority voting manner based on results of the primary vessel branch classification of the vascular unit structures belonging to the same chunk among the plurality of vascular unit structures,
wherein the segment is composed of vascular unit structures belonging to a region segmented by a branch point in the vascular structure.
7 . The method of claim 1 , wherein the classifying the vessel branch includes:
performing, by the analysis device, primary vessel branch classification for each of the plurality of vascular unit structures belonging to same chunk by using the second learning model; and
correcting, by the analysis device, the primary vessel branch classification result, when results of the primary vessel branch classification of vascular unit structures belonging to same chunk among the plurality of vascular unit structures are different from information identified in the cerebrovascular image,
wherein the segment is composed of vascular unit structures belonging to a region segmented by a branch point in the vascular structure.
8 . An analysis device for analyzing a cerebrovascular branch in cerebrovascular image, the device comprising:
an input device receiving a cerebrovascular image of a subject;
a storage device storing a first learning model that classifies chunks to which a vascular unit structure belongs and a second learning model that classifies a cerebrovascular branch of the vascular unit structure belonging to the same chunk; and
an arithmetic device extracting a plurality of vascular unit structures based on geometric features of a three-dimensional (3D) model from the cerebrovascular image, inputting feature values for each of the plurality of vascular unit structures into the first learning model to classify chunks to which each of the plurality of vascular unit structures belongs, and inputting feature values of each of the vascular unit structures belonging to the same chunk into the second learning model to analyze vessel branch characteristics constituted by the vascular unit structures belonging to the same chunk, and wherein each vascular unit structure is a spot, and the spot is a cell having a regular spacing from the artery center line extracted from the cerebrovascular image.
9 . The device of claim 8 , wherein the chunk is a vascular unit structure segmented into different types according to at least one criterion of (i) symmetry; (ii) anterior or posterior; (iii) basal or pial; and (iv) a group including criteria of middle cerebral arteries (MCA), anterior cerebral arteries (ACA) and posterior cerebral arteries (PCA), as a higher structure of a vessel branch including at least one vessel branch type.
10 . The device of claim 8 , wherein the feature values include cerebral vessel cross-sectional area, maximum inscribed sphere radius, minimum diameter, maximum diameter, maximum-minimum radius ratio, surface circumference, distortion, curvature and lumen roundness.
11 . The device of claim 10 , wherein the feature values further include a brightness value of the vascular unit structure.
12 . The device of claim 8 , wherein the arithmetic device performs primary chunk classification for each of the plurality of vascular unit structures using the first learning model, and performs secondary chunk classification for the vascular unit structures belonging to same segment, in a majority voting manner based on results of the primary chunk classification of the vascular unit structures belonging to the same segment among the plurality of vascular unit structures; and
wherein the segment is composed of vascular unit structures belonging to a region segmented by a branch point in a vascular structure.
13 . The device of claim 8 , wherein the vessel branch characteristics are vessel branch classifications of the vascular unit structure belonging to the same chunk.
14 . The device of claim 8 , wherein the vessel branch characteristics are results of identifying whether or not the subject has a brain disease.