南京理工大学学报（自然科学版）2012，Vol.36Issue(2)：332-337,6.
一种改进的最大类间方差图像分割法
Image Segmentation Based on Improved Otsu Algorithm
摘要
Abstract
According to the features of white matter in the human brain slice images ( an improved maximum between-clusler variance ( Olsu) algorithm for the image segmentation is proposed by introducing the average variance into the conventional Otsu method. The images are pre-segmented by the connected component labeling method and then filtered by means of the open operation of morphology. Finally,the improved Otsu theory is applied to extract the white matter from the slice images. The proposed algorithm is compared with the conventional Otsu algorithm in terms of their performances on the segmentation of white matter in sequential slice images qualitatively, and evaluated quantitatively by using the missed detection rate and the false detection rate. The average missed detection rates of the two methods are all 0. 03 around, and the average false declection are 0.026 and 0.251 respectively. The results indicate that the improved Otsu method characterized by the combination of the region information and edge information can lead to more accurate and effective segmentation.关键词
最大类间方差(大津法)/图像分割/阈值/白质/开运算Key words
maximum between-duster variance ( Otsu ) / image segmentation/ threshhold/ white matters/open operation分类
信息技术与安全科学引用本文复制引用
李敏,罗洪艳,郑小林,谭立文,朱文武..一种改进的最大类间方差图像分割法[J].南京理工大学学报（自然科学版）,2012,36(2):332-337,6.基金项目
国家自然科学基金(60771025 （）
30900323) （）
中央高校基本科研业务费资助项目(CDJXS10231122) （CDJXS10231122）
