Método de Parcelamento do Corpo Caloso por Watershed em Imagens de Tensor de Difusão

The corpus callosum is the most important white matter brain structure that connects both cerebral hemispheres, and represents an important structure studying for neurodegenerative diseases. Thus the corpus callosum characterization and parcellation are desirable tasks, but due to the lack of visible landmarks, shape and intensity variation among subjects, it represents a complex task. This paper aims to reproduce a previously proposed corpus callosum parcellation method for fifty subjects, analyze possible flaws and contribute with potential improvements. The method was based on the Watershed algorithm in the magnetic resonance diffusion tensor images, and experiments were performed in a limited dataset with 15 patients.