Please use this identifier to cite or link to this item: http://repositorio.unicamp.br/jspui/handle/REPOSIP/92067
Type: Artigo de evento
Title: Segmentation Of Brain Structures By Watershed Transform On Tensorial Morphological Gradient Of Diffusion Tensor Imaging
Author: Rittner L.
Appenzeller S.
Lotufo R.
Abstract: Watershed transform on tensorial morphological gradient (TMG) is a new approach to segment diffusion tensor images (DTI). Since the TMG is able to express the tensorial dissimilarities in a single scalar image, the segmentation problem of DTI is then reduced to a scalar image segmentation problem. Therefore, it can be addressed by well-known segmentation techniques, such as the watershed transform. In other words, by computing the TMG of a DTI, and then using the hierarchical watershed transform, it is possible to segment brain structures, such as the corpus callosum, the ventricles and the cortico-spinal tracts, and use the results for subsequent quantitative analysis of DTI parameters. Experiments showed that segmentations obtained with the proposed approach are similar to the ones obtained by other segmentation techniques based on DTI and also segmentation methods based on other Magnetic Resonance Imaging (MRI) modalities. Since the proposed method, as opposed to the majority of the DTI based segmentation methods, does not require manual seed and/or surface placement, its results are highly repeatable. And unlike other methods that have sometimes four parameters to be adjusted, the only adjustable parameter is the number of regions in which the image should be segmented, making it simple and robust. © 2009 IEEE.
Editor: 
Rights: fechado
Identifier DOI: 10.1109/SIBGRAPI.2009.36
Address: http://www.scopus.com/inward/record.url?eid=2-s2.0-77949639597&partnerID=40&md5=1b32556bc8b2a67eacef5c5dde5093b7
Date Issue: 2009
Appears in Collections:Unicamp - Artigos e Outros Documentos

Files in This Item:
File Description SizeFormat 
2-s2.0-77949639597.pdf523.9 kBAdobe PDFView/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.