Multimodal MRI image segmentation method for glioblastoma

Abstract

The present invention proposes a multimodal nuclear magnetic resonance image segmentation method for glioblastoma, which adopts a segmentation strategy combining random forest and region growing method to segment the results of multimodal magnetic resonance images of brain glioblastoma by region growth Replace the random forest rough segmentation results with low confidence, generate retraining data and train the random forest model again, finely segment the multimodal magnetic resonance images of glioma, and segment the brain magnetic resonance images into normal brain tissue areas, necrotic areas There are 5 parts: area, active tumor area, T1 abnormal area and FLAIR abnormal area. The present invention aims at fine segmentation and positioning of glioblastoma, and assists doctors in diagnosis and treatment, so that glioblastoma can be accurately located and different sub-regions of the tumor can be segmented more accurately and finely, which is convenient for doctors to be faster and more accurate Accurately diagnose glioblastoma and formulate an accurate treatment plan.

Description

technical field [0001] The invention relates to the fields of digital medical image analysis and intelligent health management, in particular to a multimodal nuclear magnetic resonance image segmentation method for glioblastoma. Background technique [0002] Glioma is the most common primary brain tumor, more than half of which are the most malignant glioblastoma (Glioblastoma Multiforme, GBM). For GBM, no matter its form is benign or malignant, it will increase intracranial pressure, compress brain tissue, cause damage to the central nervous system, and endanger the life of the patient. The location and quantitative calculation of GBM lesion tissue (such as tumor volume, etc.) are crucial for the diagnosis, surgical planning, and postoperative analysis of GBM. In clinical practice, radiologists usually manually segment tumors from multimodal magnetic resonance images, which is tedious and time-consuming. Automatically segmenting GBM tumors can not only free doctors from i...

AI Technical Summary

Problems solved by technology

However, through the analysis of the existing GBM multimodal magnetic resonance image segmentation methods, it is found that although the above studies have achieved certain research results, they all focus on searching and segmenting the entire tumor area, and do not further subdivide the entire tumor into tumors. sub-region

Images