CT brain parenchyma segmentation system, method and device based on block region growing method

Abstract

The invention discloses a CT brain parenchyma segmentation system based on a block region growing method, which comprises a preprocessing module, a coarse segmentation module, a brain parenchyma extraction module and a post-processing module. The preprocessing module is used for preprocessing original three-dimensional CT image data to obtain preprocessed data; the coarse segmentation module is used for segmenting the preprocessed data by adopting a threshold segmentation method to obtain a foreground and an initial skull part, and processing the initial skull part by adopting binary morphology to obtain a brain parenchyma mask; the brain parenchyma extraction module is used for acquiring seed points of a brain parenchyma region from the brain parenchyma mask, and performing block circulation segmentation through a three-dimensional region growing algorithm to obtain a whole brain parenchyma mask; and the post-processing module subtracts the initial skull part from the whole brain parenchyma mask to obtain a final brain parenchyma segmentation result. The idea of partitioning is applied to the region growing segmentation method, and rapid and stable segmentation of brain parenchyma is achieved.

Description

technical field [0001] The invention relates to the technical field of cranial CT image processing, in particular to a CT brain parenchymal segmentation system, method, device and medium based on a block region growing method. Background technique [0002] Automated skull stripping or brain extraction is the process of isolating brain tissue on radiological images. This technique has become a valuable image processing tool in neuroimaging research. The literature on brain extraction tools for magnetic resonance imaging (MRI) is well established, and the Brain Extraction Tool (BET) in the FMRIB Software Library (FSL) is a reliable and validated brain extraction tool. However, gaps exist in research on brain extraction for computed tomography (CT). Especially distressing is the fact that CT is more widely used in clinical practice for the diagnosis and treatment of stroke patients than MRI. CT image postprocessing is a necessary step for ischemic stroke lesion and hemorrhag...

AI Technical Summary

Problems solved by technology

However, the method was only validated on 22 NCCT images and no results were reported on CTA images

Also, like any other deep learning based segmentation method, it requires a large number of properly labeled training images, which is time and labor intensive

Images