Improving Lesion Classification Using an Empirical Knowledge of False Classifications in Multiple Sclerosis

Keywords

accuracy accurate acknowledgment acquisition aligned aligning anatomical appear application applied arrow atlases attenuated automated axial body brain cerebellar cerebellum chart class classification classifications classified classifier cohort combined connectivity consistent consists cortical created custom deep deformation deformed delineated developed diagnostic diagram dice dual eliminate empirical equal expert exploited exploiting extent extract extraction false fields fifty filtration finally flair flow fluid fuzzy generally generated gradient gray greater health hence histogram improve improved improvements improving intensity inverse isolated isotropic knowledge known labeling lesion lesions linear load located mask masking mimic minimize minimized minimizing necessary noise optimal paired parametric part prepared presence prevalent probabilities processed proposed quasar rapid realized recovery reformatted registration remaining removed representing respectively rest retrospectively rigid saturation scanner schematic scheme sclerosis segmentation segmented separately seven significantly similarity since software space structures subject subjects subsequently subtracting supported suppressing symmetric system tailed taking template threshold thresholds tissues tracking uniformity utilized validated validation volumetric whole