An Automatic Method Of Brain Extraction Based On Active ROI

Automatic extraction of brain is an important step in the preprocessing of brain analyses. Subsequent analyses,such as brain tissues segmentation,measurement of brain volume and brain atrophy are largely influenced by the results of brain extraction.The brain extraction is divided into manual, semiautomatic and automatic methods according to the artificial participation.The manual method is time-consuming and required the operator with professional ability while the automatic method make the extraction a result with less time consuming as the apply of new algorithms.So the automatic method gets great attentions of researchers in recent years.This paper presents an automatic method of brain extraction based on AROI(active region of interested).This paper first introduces the background and introdudces BET which is a commonly used algorithm and the improved BET algorithm, then describes the GAC model based on edge information and the CV model based on regional information.At last, this paper introduces a hybrid model which combines the edge and the region information.This paper presents an automatic method of brain extraction based on AROI. The first step of the proposed algorithm is processing the MRI with the improved BET algorithm which can automaticlly converges to the rough outline of brain quickly. After that,we make the morphological expansion processing on the rough outline of brain,so we get the initial region of interested(ROI).An improved hybrid active contour model then be used in the ROI,we can get a new contour,and after expansion on the new contour.we get the new ROI.So the hybrid contour model make the evolutation of the brain contour in the actived (continuously updated) ROI until the hybrid model converged. Finally we get the a more precise contour of brain.