Evaluation of morphological reconstruction, fast marching and a novel hybrid segmentation method

Authors:
Jianfeng Xu;Lixu Gu
Affiliations:
Computer Science, Shanghai Jiaotong University, Shanghai, P.R. China;Computer Science, Shanghai Jiaotong University, Shanghai, P.R. China
Venue:
CIS'04 Proceedings of the First international conference on Computational and Information Science
Year:
2004

Citing 3
Cited 1

Fronts propagating with curvature-dependent speed: algorithms based on Hamilton-Jacobi formulations

Journal of Computational Physics
Computing minimal surfaces via level set curvature flow

Journal of Computational Physics
Fast Marching Methods

SIAM Review

Medical image alignment by normal vector information

CIS'05 Proceedings of the 2005 international conference on Computational Intelligence and Security - Volume Part I

Quantified Score

Hi-index	0.00

Visualization

Abstract

An evaluation of two traditional segmentation algorithms of Morphological Reconstruction and the Fast Marching method along with a novel hybrid segmentation approach is presented. After introducing the Fast Marching and the Morphological Reconstruction segmentation, we propose a novel hybrid segmentation approach in multi-stage, which is derived from both an improved Fast Marching method and the Morphological Reconstruction. To demonstrate the effectiveness and accuracy of the three methods, we employ an MRI brain image in our experiments, in which “gold standard” is known. The evaluation is measured accordingly in accuracy and speed when running a 2.0 GHz based windows XP PC. The accuracy results of average 0.9738, 0.6302 and 0.9734 measured in similarity indexes of the Morphological Reconstruction, the Fast Marching and the hybrid approach are achieved, respectively. The computing performance required 188.6, 22.3 and 43.4 in seconds accordingly.