直方图的脑图像分割策略

Histogram-Based Method for Magnetic Resonance Brain Image Segmentation

为了提高核磁共振分割精度和效率，提出了一种基于直方图核磁共振脑图像分割算法。利用脑提取工具剔除脑核磁共振图像中非脑组织。根据灰度变化的特点，将脑图像直方图的区域峰值组成新的直方图，通过基于直方图阈值选取规则将脑组织分割成白质、灰质、脑脊液。实验结果表明：该算法相对于常见的分割算法，获得了较高的分割精度，同时运算效率高，运算时间降低了1～2个数量级。

This paper presents a histogram-based segmentation method for magnetic resonance brain images in order to improve segmentation accuracy and efficiency .Brain Extraction Tool method is used to remove non-brain tissues .According to the characteristics of grayscale change ,a new histogram is extracted from the brain image histogram peaks ,and Brain tissue was segmented into white matter （WM ） ,gray matter （GM ） and cerebrospinal fluid （CSF） by using the new histogram threshold rule . The experimental results show that ,compared with the common segmentation algorithm , the new method is of higher segmentation accuracy and computational efficiency , and the computing time is reduced by 1 to 2 orders of magnitude .