Spatio-temporal segmentation for radiotherapy planning
J. Stawiaski (Centre de Morphologie Mathématique-Ecoles des Mines de Paris, France) and E. Decencière (Centre de Morphologie Mathématique-Ecoles des Mines de Paris, France)
This paper presents a segmentation method for spatio-temporal data (4D images) for radiotherapy planning. The aim of this study is to propose some techniques for the segmentation of tumours surrounding or contained in the lungs. The 4D images are produced using a respiration gating procedure and computed tomography. The aim of the segmentation is to follow the tumour movement while the patient is breathing, so that he does not need to hold his respiration during the radiation treatment.
The proposed technique is based on mathematical morphology.
It uses a 4D watershed algorithm, combined with graph-based techniques. The differences between different classical spatio-temporal segmentation algorithms will be highlighted, and conclusions on the related trade-offs between speed and precision will be drawn.