Mathematical Morphology Application to Automatic Segmentation of Extraocular Muscles in Two-dimensional CT

André Domingos Araújo Souza, MSc

Imagcom-Dept. Physics and Mathematics

University of São Paulo-Ribeirão Preto, Brazil

PURPOSE:

Extraocular muscle (EOM) measurements are usually performed on 2D CT or MRI images. Typically this measurement requires a subjective, inaccurate and time consuming identification of the EOMs. A new algorithm, is based entirely in mathematical morphology, that quantify and isolate EOM automatically is presented