Accurate segmentation of the right ventricle (RV) is a crucial step in the assessment of the ventricular structure and function. Yet, due to its complex anatomy and motion segmentation of the RV has not been as largely studied as the left ventricle. This paper presents a fully automatic method for the segmentation of the RV in cardiac magnetic resonance images (MRI). The method uses a coarse-to-fine segmentation strategy in combination with a multi-atlas propagation segmentation framework. Based on a cross correlation metric, our method selects the best atlases for propagation allowing the refinement of the segmentation at each iteration of the propagation. The proposed method was evaluated on 32 cardiac MRI datasets provided by the RV Segmentation Challenge in Cardiac MRI.