Evaluation of a thoracic elastic registration method using anatomical constraints in oncology

The aim of our work is to improve data analysis in thoracic oncology applications by means of non-rigid registration of CT and PET images. A mutual information-driven transformation using B-spline free-form deformations performs the alignment of the image volumes. Special constraints relying on hierarchically identifying corresponding structures in both modalities have been added to guarantee the convergence towards an optimal registration.

[1]  H N Wagner.  SNM 1999. Fused image tomography: an integrating force. , 1999, Journal of nuclear medicine : official publication, Society of Nuclear Medicine.

[2]  Paul A. Viola,et al.  Alignment by Maximization of Mutual Information , 1995, Proceedings of IEEE International Conference on Computer Vision.

[3]  T. Géraud.  Segmentation des structures internes du cerveau en imagerie par resonance magnetique tridimensionnelle , 1998 .

[4]  Daniel Rueckert,et al.  Nonrigid registration using free-form deformations: application to breast MR images , 1999, IEEE Transactions on Medical Imaging.