The competition between discrete (MRF based) and continuous (PDE based) formulations has a very long history, especially in context of segmentation. Obviously, both have their advantages and drawbacks. Therefore the choice of a discrete or continuous framework is often driven by a particular application or (even more often) by personal preferences of a particular researcher. In this work we present a model for binary segmentation, where discrete and continuous parts are combined in a well founded and simple way. We discuss the properties of the proposed model, give a basic inference algorithm and validate it on a benchmark database. © 2013 Springer-Verlag.
CITATION STYLE
Schlesinger, D. (2013). A continuous shape prior for MRF-based segmentation. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 8081 LNCS, pp. 350–361). https://doi.org/10.1007/978-3-642-40395-8_26
Mendeley helps you to discover research relevant for your work.