9 May 2002 Automated quantification of MS lesions in MRI: a validation study
Author Affiliations +
Abstract
Two novel methods for automated quantification of total lesion burden in multiple sclerosis patients using multi-spectral magnetic resonance (MR) imaging are examined. The first method, geometrically constrained region growth, requires user specification of lesion location. The second, directed multi-spectral segmentation, requires only the location of a single exemplar lesion. The performances of these methods are compared to manual tracing using three parameters: speed, precision, and accuracy. Both methods are shown to provide significant improvement over manual tracing in terms of processing time, inter- and intra-operator coefficients of variation, and global accuracy using both phantoms and clinical data.
© (2002) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Edward A. Ashton, Edward A. Ashton, Chihiro Takahashi, Chihiro Takahashi, Michel J. Berg, Michel J. Berg, Andrew Goodman, Andrew Goodman, Saara Totterman, Saara Totterman, } "Automated quantification of MS lesions in MRI: a validation study", Proc. SPIE 4684, Medical Imaging 2002: Image Processing, (9 May 2002); doi: 10.1117/12.467157; https://doi.org/10.1117/12.467157
PROCEEDINGS
9 PAGES


SHARE
Back to Top