A method for semantic segmentation of 3D liver images based on contextual attention strategy

Abstract

The present invention relates to a method for semantic segmentation of three-dimensional liver images based on a contextual attention strategy, comprising the following steps: selecting a medical image data set to be segmented into a liver, and dividing it into a training set and a test set; The 3D liver image is preprocessed; in the encoding stage, the feature map of the liver is obtained by using the residual structure, convolutional network and atrous convolution; in the decoding stage, the liver is obtained by using the context attention strategy module, transposed convolution and depth supervision The segmented image of the image; the post-processing of the liver image obtained after the semantic segmentation. This method has the characteristics of improving the semantic segmentation effect of 3D liver images, achieves a better automatic segmentation effect, and can assist doctors in diagnosis.

Description

technical field [0001] The invention relates to a three-dimensional medical image semantic segmentation method, in particular to a three-dimensional liver image semantic segmentation method based on a context attention strategy. Background technique [0002] The liver is located in the abdomen of the human body and is the largest important solid organ in the abdomen. However, liver-related diseases such as liver cancer have become one of the most common diseases with the highest mortality rate in the world, which poses a great threat to human health and life. . In recent years, computed tomography (CT) has become the most widely used medical imaging method for the discovery, diagnosis and treatment of liver tumors. Detailed understanding of the shape and position of the liver in CT images is required before the treatment and operation, so the accurate segmentation of the liver has become the primary task of liver cancer treatment. However, the size, shape and location of t...

AI Technical Summary

Problems solved by technology

However, the two-dimensional convolutional network cannot make full use of the spatial information in the medical image data, and lacks the processing of the information between the slices of the three-dimensional medical data, so the boundary of the segmentation result is relatively rough, and the overall effect of the segmentation is not as good as that of the three-dimensional segmentation.

However, these studies have not processed the information between slices of medical image data, and the fusion of low-level semantic features and high-level semantic features of the target is not fully utilized.

Images