Share Email Print

Proceedings Paper

Towards an automatic tumor segmentation using iterative watersheds
Author(s): Matei Mancas; Bernard Gosselin
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

This paper introduces a simple knowledge model on CT (Computed Tomography) images which provides high level information. A novel method called iterative watersheds is then used in order to segment the tumors. Moreover, a fully automatic tumor segmentation method was tested by using image registration. Some preliminary results are very encouraging and give us hope to obtain an interesting tool for the clinic. Tests were made on head and neck images, nevertheless, this is a generic method working on all kinds of tumors. The iterative watersheds and our model are first introduced, then PET (Positron Emission Tomography) images registration on CT is described. Some results of iterative watersheds are compared using either the semi-automatic or fully automatic mode. Finally we conclude by a discussion about operator's interaction and important future work.

Paper Details

Date Published: 12 May 2004
PDF: 11 pages
Proc. SPIE 5370, Medical Imaging 2004: Image Processing, (12 May 2004); doi: 10.1117/12.535017
Show Author Affiliations
Matei Mancas, Faculte Polytechnique de Mons (Belgium)
Bernard Gosselin, Faculte Polytechnique de Mons (Belgium)

Published in SPIE Proceedings Vol. 5370:
Medical Imaging 2004: Image Processing
J. Michael Fitzpatrick; Milan Sonka, Editor(s)

© SPIE. Terms of Use
Back to Top