CT image lung vessel segmentation method

Abstract

The invention relates to a CT image lung vessel segmentation method. The method comprises the steps of S1, acquiring a to-be-segmented image; s2, performing data preprocessing on the to-be-segmented image; s3, sampling a local input image block; s4, inputting the image blocks into a pre-trained segmentation model; s5, post-processing a segmentation result, and constructing a blood vessel tree diagram for refinement; and S6, carrying out three-dimensional visualization on a segmentation result. A segmentation network model is an improved 3D Unet model, and a multi-scale strategy is added to theimproved 3D Unet model to capture rich context information so as to improve the characteristic performance of pulmonary vessels. An attention mechanism is introduced for pulmonary vessel segmentationof pathological lung data, a related region of a target can be highlighted adaptively, and irrelevant features, especially the influence of a lesion region on vessel segmentation, can be inhibited implicitly.

Description

technical field [0001] The invention belongs to the fields of computer vision and digital image processing, and in particular relates to a method for segmenting pulmonary vessels in CT images. Background technique [0002] Pulmonary vessel segmentation is indispensable in computer-aided detection systems. Accurate pulmonary vessel segmentation plays an important role in clinical diagnosis, analysis of lesion and surgical planning. Using imaging, digital image processing and computer vision technology in the computer-aided detection system, the pulmonary vascular structure in the lung CT image is automatically segmented and visualized in 3D, so as to obtain the characteristic information of the target tissue structure, which can provide doctors with more information. The observation method gives doctors more references and helps doctors make faster and more accurate judgments. [0003] Traditional pulmonary vessel segmentation methods mostly rely on handcrafted features, re...

AI Technical Summary

Problems solved by technology

[0004] As mentioned above, the currently commonly used pulmonary vessel segmentation method based on blood vessel enhancement requires a given threshold to distinguish pulmonary vessels from background features. The selection of the threshold has high requirements, and improper selection is prone to missing segmentation or over-segmentation; based on edge and region The deformable model needs to design the objective function of the task, and obtain the minimum energy contour under the action of internal and external forces. In complex cases, the amount of calculation is too large and it is difficult to converge; based on tracking algorithms, manual setting of seed points is generally used, which is time-consuming and difficult. Automate

In the application of traditional methods in the segmentation task of pulmonary vessels, the generalization and robustness of the method are poor, and good results can only be achieved in specific environments.

In the deep learning segmentation method, only for the segmentation of health data, the results that cannot meet the needs of doctors in actual clinical use

Images