一种脑切片图像分割新法
    点此下载全文
引用本文:侯发忠,邹北骥,刘召斌,周支元.一种脑切片图像分割新法[J].计算技术与自动化,2015,(4):90-94
摘要点击次数: 1417
全文下载次数: 109
作者单位
侯发忠,邹北骥,刘召斌,周支元 (1. 湖南医药学院, 湖南 怀化418000
2.中南大学 信息科学与工程学院, 湖南 长沙410083) 
中文摘要:针对脑切片的特点和普通分割方法的不足,提出一种基于改进的Otsu分割方法,在Otsu分类后的交叉区域,预先指定若干个前、背景点,再以这些关键点的阈值进行最小类间方差分割,同时用关键点的阈值建立等高线,根据相邻切片图像的组织阈值具有相似性,依次将单张切片的分割结果作为相邻切片分割的初始值,进而实现整个数据集中的组织连续自动分割。结果表明:该方法能准确从彩色人脑切片图像中分割提取出复杂的器官组织,克服了其它方法的缺陷和交互式方法的人工依赖,并提高了分割的精度。
中文关键词:切片图像  Otsu算法  多层分割;三维重建
 
New Segmentation of Slice Images of Digital Human Brain
Abstract:Based on the characteristics of brain slice images and aiming at the shortcomings of segmentation methods, a segmentation method based on improved Otsu was proposed. In the crossover region after classification, several points of target and background were specified beforehand, the threshold values of the key points were divided into the minimum class variance and the contour lines were established at the same time. According to the similarities of the threshold value of organ of adjacent slice images, the segmented results of the single slice were used as the initial value of the next one to realize the continuous and automatic segmentation of organ in brain slice image sequences.The experimental results indicate that this method can extract complex organ from color brain slice images accurately, reduce the massive manual intervention involved and the defects in existing segmentation methods, and improve segmentation accuracy.
keywords:slice images  Otsu algorithm  multilayer segmentation  3D reconstruction
查看全文   查看/发表评论   下载pdf阅读器