10.1371/journal.pone.0198335.t002 Omer Faruk Gulban Omer Faruk Gulban Marian Schneider Marian Schneider Ingo Marquardt Ingo Marquardt Roy A. M. Haast Roy A. M. Haast Federico De Martino Federico De Martino Segmentation performance scores MPRAGE data set. Public Library of Science 2018 post-processing step aids future efforts voxel matter borders 7 Tesla MRI data approach data analysis Python implementation field MRI High-resolution matter ribbon dura mater ground truth blood vessels data slice-by-slice error scale properties change scalable method matter segmentation histogram representations Residual segmentation errors test segmentation algorithms resonance imaging 2018-06-06 17:46:16 Dataset https://plos.figshare.com/articles/dataset/Segmentation_performance_scores_MPRAGE_data_set_/6452150 <p>The table shows the DICE (larger is better) and AVHD (less is better) for the initial SPM 12 and FSL FAST GM segmentations as well as after additional polishing, using either the gradient magnitude or the compositional data method.</p>