Automated medical image segmentation using a new deformable surface model

Title: Automated medical image segmentation using a new deformable surface model
Creator: Luo, Suhuai
Relation: International Journal of Computer Science and Network Security Vol. 6, Issue 5A, p. 109-115
Relation: http://search.ijcsns.org/02_search/02_search_03.php?number=200605016
Publisher: IJCSNS
Resource Type: journal article
Date: 2006
Description: This paper introduces an automated medical image segmentation algorithm which can be used to locate volumetric objects such as brain tumor in Magnetic Resonance Imaging (MRI) images. The algorithm is novel in that it deals with MRI slices (or images) as a three dimension (3D) object as a whole. All the processes of segmentation are done in 3D space. First, it removes noisy voxels with 3D nonlinear anisotropic filtering. The filtering well preserves the intensity distribution continuity in all three directions as well as smoothes noisy voxels. Second, it uses a novel deformable surface model to segment an object from the MRI. A dynamic gradient vector flow is used in forming the surface model. Experiments have been done on segmenting tumors from real MRI data of human head. Accurate 3D tumor segmentation has been achieved.
Subject: segmentation; anisotropic filtering; deformable surface model
Identifier: uon:964
Identifier: http://hdl.handle.net/1959.13/26601
Identifier: ISSN:1738-7906
Reviewed

		Thumbnail	File	Description	Size	Format
View Details Download			ATTACHMENT01	Publisher version (open access)	533 KB	Adobe Acrobat PDF	View Details Download