A Novel and Multi-Scale Unsupervised Algorithm for Image Segmentation

摘要：

Gibbs Random Fields (GRF) is apopular prior model widely used in Bayesiansegmentation due to its excellent property indescribing the spatial information of image. Butuntil now, the classical approaches, describing theMarkovian property of single-scale instead that ofmulti-scale, may come across some difficultiessuch as expensive computation and unsupervisedparameter estimation of GRF. Thus, in this paper,a novel and unsupervised algorithm namedmulti-scale GRF that addresses these problemsperfectly is proposed by extending the classicalsingle-scale model of GRF to a multi-scale one atthe first time. Experiments have shown that ouralgorithm presented in the paper has excellentrobustness and easy to be used in unsupervisedand precise segmentation.

作者: Luo Minmin Jiang Guiping Lin Ya-zhong

作者单位: School of Biomedical Engineering Southern Medical University Guangzhou China

会议类型: 国际会议

会议名称: The 4th International Conference on Bioinformatics and Biomedical Engineering(第四届IEEE生物信息与生物医学工程国际会议 iCBBE 2010)

会议地点: 成都

会议语种:英文

页码: 1-5

在线出版日期: 2010-06-18（万方平台首次上网日期，不代表论文的发表时间）