Discrete Multi Atlas Segmentation using Agreement Constraints

Stavros Alchatzidis, Aristeidis Sotiras and Nikos Paragios

In Proceedings British Machine Vision Conference 2014


In this paper we propose an approach for integrated volume segmentation and multi-atlas registration. The aim is to recover simultaneously all atlases deformations along with a segmentation mask that imposes consistency between their deformed segmentation maps. This is modeled through a pairwise graphical model where deformation variables are coupled with the segmentation ones. Dual decomposition is used to recover the optimal solution of the method. Promising results on the IBSR dataset demonstrate the potential of our method where improvement is obtained when comparing to the post-registration label fusion.


Segmentation and Object Detection


Extended Abstract (PDF, 1 page, 192K)
Paper (PDF, 12 pages, 544K)
Bibtex File



Stavros Alchatzidis, Aristeidis Sotiras, and Nikos Paragios. Discrete Multi Atlas Segmentation using Agreement Constraints. Proceedings of the British Machine Vision Conference. BMVA Press, September 2014.


	title = {Discrete Multi Atlas Segmentation using Agreement Constraints},
	author = {Alchatzidis, Stavros and Sotiras, Aristeidis and Paragios, Nikos},
	year = {2014},
	booktitle = {Proceedings of the British Machine Vision Conference},
	publisher = {BMVA Press},
	editors = {Valstar, Michel and French, Andrew and Pridmore, Tony}
	doi = { http://dx.doi.org/10.5244/C.28.20 }