Improved 3D U-Net model-based preschool child lung image region-of-interest segmentation method

Abstract

The invention provides a preschool child lung image region-of-interest segmentation method based on an improved 3D U-Net model. The preschool child lung image region-of-interest segmentation method comprises the following steps: (1) collecting CT image data of a preschool child patient for preprocessing; (2) dividing the preprocessed image into a training set, a verification set and a test set; (3) a segmentation model is constructed, the segmentation model adopts an improved 3D U-Net network model, a channelized Transform module is designed between an encoder and a decoder of the 3D U-Net network model, and a UCTransNet framework is constructed to replace jump connection in U-Net so as to better fuse characteristics of the encoder; (4) sending the preprocessed training set into the constructed segmentation model for training; and (5) inputting a to-be-segmented lung image of the preschool child into the trained segmentation model to obtain a region of interest of the lung image. According to the method, the problem that the image quality is lower than that of adults due to hyperactivity during CT shooting of preschool children can be relieved, so that the lung region-of-interest is automatically and accurately segmented from the CT image.

Description

technical field [0001] The invention belongs to the field of medical artificial intelligence, and in particular relates to a method for segmenting a region of interest in preschool children's lung images based on an improved 3D U-Net model. Background technique [0002] Computer vision technology is often used in the field of fast and intelligent image processing, such as image classification, target detection and target retrieval. Computer vision simulates the human visual mechanism and has the advantages of fast detection speed and low cost. [0003] In recent years, with the application of deep learning in the field of computer vision, especially in the field of medical imaging, breakthroughs have been made. The data-driven deep learning technology allows computers to combine imaging and medical image processing technology with computer technology. Analysis and calculation can realize automatic segmentation of target areas. The segmentation of organ tissues or lesions in ...

AI Technical Summary

Problems solved by technology

[0007] The invention provides an improved 3D U-Net model-based method for segmenting regions of interest in preschool children's lung images, which can alleviate the problem that the image quality of preschool children is not as good as that of adults due to hyperactivity when taking CT images, thereby automatically and Accurately segment lung regions of interest

Images