Skip to main content
. Author manuscript; available in PMC: 2016 Dec 1.
Published in final edited form as: Med Image Anal. 2015 Aug 28;26(1):82–91. doi: 10.1016/j.media.2015.08.010

Figure 8.

Figure 8

Results of empirical evaluation. The results indicate: without using non-local information, the MLF framework (large-scale) provides better performance than two multi-atlas segmentation algorithms (majority vote and SS) and has comparable performance as the JLF benchmark. Note that, the multi-atlas segmentation used “non-rigid registration + fusion” framework while the JLF and the MLF used “affine registration + fusion” framework.