Medical image segmentation method based on cross-scale correction consistency learning

• Main Authors: DONG MIN, TANG XUTAO, LI DEZHEN, LI JUNJIE, YANG ATING, LI YONGWEN, JIAO ZHANWEI, WANG ZHENXING
• Format: Patent
• Language: Chinese; English
• Published: 19.04.2024
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; IMAGE DATA PROCESSING OR GENERATION, IN GENERAL; PHYSICS

The invention belongs to the field of medical image processing, and particularly relates to a medical image segmentation method based on a cross-scale correction consistency learning algorithm, and the method comprises the steps: data preprocessing: employing a PARSE pulmonary artery public data set, dividing the data set into a training set and a test set according to a proportion, and carrying out the preprocessing of the data set, the method comprises the following steps: firstly, normalizing the pixel intensity of an image into a zero mean value and a unit variance, and then cutting out a region of interest for subsequent training; a medical image segmentation model based on cross-scale correction consistency learning is constructed, multiple views are created by adopting different enhancement modes in an input layer, a pyramid structure is introduced in an output layer to generate multi-scale output, and image-level representation is expanded; and training the network model, and sending the test set into