Brain tumor image segmentation method based on HDC-Net0 network

Abstract

As is well known, malignant tumors are globally diseases which are difficult to treat and become one of the main diseases harming human life at present. Early accurate diagnosis, positioning, qualification, classification and segmentation of tumors are the key of follow-up treatment. The MRI brain tumor image segmentation means that a whole tumor area, a tumor core area and a tumor enhancement area are segmented from normal brain tissue. A traditional segmentation algorithm mainly challenges large gray similarity between brain tissues in an MRI image and differences between different cases. According to the multi-modal MRI brain tumor image segmentation, feature information of different modals in the MRI image can be fully utilized, the segmentation effectiveness is improved, and the multi-modal MRI brain tumor image segmentation is a research hotspot of brain tumor image processing in recent years.

Description

technical field [0001] The invention belongs to the technical field of medical image processing, and in particular relates to a segmentation method for brain tumor images. Background technique [0002] Brain tumor is one of the most common and extremely harmful diseases to the human body, with high morbidity and mortality. It has a strong social hazard. According to the survey, the number of brain tumor patients in my country is increasing year by year, and the situation is very serious. The magnetic resonance image analysis of brain tumors is an important basis for doctors to diagnose and treat brain tumors, evaluate surgical procedures, and track the condition. However, brain tumors have various shapes and complex structures, and manual segmentation is very time-consuming and labor-intensive, and it is easy to misjudgment during the segmentation process. Therefore, combining computer-aided diagnosis with computer vision technology, a fully automatic and precise segmentat...

AI Technical Summary

Problems solved by technology

[0004] The content of the present invention is to solve the insufficient segmentation accuracy of brain tumor images and provide an automatic and accurate method for segmenting brain tumor images

[0006] (1) Aiming at the problems that the current 3D convolutional network has large computing memory and low computing efficiency, a lightweight pseudo-3D hierarchical decoupling network (HDC-Net 0 ) model, which completes the 3D segmentation task of brain tumors based on 2D convolution, and achieves excellent segmentation accuracy, which can effectively solve the problem of limited computing overhead and GPU memory consumption when performing brain tumor segmentation tasks based on 3D networks

Images