14 April 2005 Hierarchical brain tissue segmentation and its application in multiple sclerosis and Alzheimer's disease
Author Affiliations +
Proceedings Volume 5746, Medical Imaging 2005: Physiology, Function, and Structure from Medical Images; (2005); doi: 10.1117/12.596297
Event: Medical Imaging, 2005, San Diego, California, United States
Abstract
Based on Fuzzy Connectedness (FC) object delineation principles and algorithms, a hierarchical brain tissue segmentation technique has been developed for MR images. After MR image background intensity inhomogeneity correction and intensity standardization, three FC objects for cerebrospinal fluid (CSF), gray matter (GM), and white matter (WM) are generated via FC object delineation, and an intracranial (IC) mask is created via morphological operations. Then, the IC mask is decomposed into parenchymal (BP) and CSF masks, while the BP mask is separated into WM and GM masks. WM mask is further divided into pure and dirty white matter masks (PWM and DWM). In Multiple Sclerosis studies, a severe white matter lesion (LS) mask is defined from DWM mask. Based on the segmented brain tissue images, a histogram-based method has been developed to find disease-specific, image-based quantitative markers for characterizing the macromolecular manifestation of the two diseases. These same procedures have been applied to 65 MS (46 patients and 19 normal subjects) and 25 AD (15 patients and 10 normal subjects) data sets, each of which consists of FSE PD- and T2-weighted MR images. Histograms representing standardized PD and T2 intensity distributions and their numerical parameters provide an effective means for characterizing the two diseases. The procedures are systematic, nearly automated, robust, and the results are reproducible.
© (2005) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Tianhu Lei, Jayaram K. Udupa, Gul Moonis, Eric Schwartz, Laura Balcer, "Hierarchical brain tissue segmentation and its application in multiple sclerosis and Alzheimer's disease", Proc. SPIE 5746, Medical Imaging 2005: Physiology, Function, and Structure from Medical Images, (14 April 2005); doi: 10.1117/12.596297; https://doi.org/10.1117/12.596297
PROCEEDINGS
12 PAGES


SHARE
KEYWORDS
Photomasks

Image segmentation

Magnetic resonance imaging

Tissues

Brain

Neuroimaging

Algorithm development

Back to Top