A method for extracting coronary arteries

Abstract

The invention provides a coronary artery extraction method. The method includes the following steps that (1) a heart three-dimensional image is input, rectification is conducted on the heart three-dimensional image and a coronary artery model, the initial point, final point and advancing direction of the coronary artery in the heart three-dimensional image are determined; (2) according to the initial point and the advancing direction of the coronary artery, the minimal cost path is initiated; (3) according to the advancing direction, if the minimal cost path advances to the final point of the coronary artery model, the minimal cost path is searched for from the final point in a reverse mode, and a center line of the coronary artery is generated; if the minimal cost path does not reach the final point, the minimal cost path is optimized and advances continuously till reaching the final point, then the minimal cost path is searched for again from the final point in a reverse mode, and the center line of the coronary artery is generated. By means of the technical scheme, the center line of the coronary artery can be extracted fully automatically, and names of coronary artery branches can be marked as well.

Description

technical field [0001] The invention relates to the field of image processing, in particular to a method for extracting coronary arteries. Background technique [0002] In clinical applications, angiographic images are one of the main means for doctors to analyze cardiovascular diseases, so the extraction of coronary artery centerline is of great significance. [0003] The traditional coronary artery centerline extraction method is mainly based on the minimum cost path (please refer to [Desca-mps, T. and L. Cohen.: In: ECCV2000.vol.1843, 543-557. (2000)]), the starting point Obtained by manual annotation or automatic detection algorithm. The energy function E(C) of the minimum cost path method on path C can be expressed as: [0004] E(C)=∫ Ω (P(C(s))+ω)ds (1) [0005] Wherein, E(C) is the energy function on the minimum cost path C; s is the arc length parameter; x=C(s), represents a certain position on C(s), and P(x) is the Energy or cost at position x on the minimum co...

AI Technical Summary

Problems solved by technology

[0013] However, after the initial point in the above method is obtained by the automatic detection algorithm, although the coronary artery can be fully automatically extracted and the coronary artery tree can be generated, it cannot automatically label the specific main branch, and the labeling process requires human-computer interaction.

In addition, for a specific main branch, this method needs to manually mark the starting point and end point, which cannot achieve the purpose of fully automatic extraction of a specific main branch

Images