Liver Contour Eztraction using Snake and Initial Boundary Auto-generation
Contour extraction of liver tissues in CT image is particularly challenging due to the anatomic complexity. An integrated model based statistical learning and active contour scheme—modified snake are presented in the paper to simplify the automatically liver contour extraction and then achieve refinements of liver boundary with accuracy. The proposed scheme consists of two sub-routines: initial contour acquisition and refinement. The former firstly extracts coarser liver regions based on estimated mixture Gaussian distribution model by the EM algorithm and then take initial contour extraction. The latter makes further refinement using modified snake algorithm with additional intensity item in external energy. Experimental results show the ability of the proposed algorithm to achieve satisfied liver boundaries in presence of liver tumor and other anatomic organs, and suggest its suitability to other medical image contour detection tasks.
EM algorithm Snake model Liver contour eztraction
Li Ma Lei Zhu
School of automation, Hangzhou Dianzi University, Hangzhou, 310018, P.R. China Dept. of computer science, Kentucky University, Lexington, KY, U.S.A
国际会议
上海
英文
2691-2694
2008-05-16（万方平台首次上网日期，不代表论文的发表时间）