|
Applied Math Seminar
Segmentation of CT Images for Radiation Treatment Planning
|
|
The process of planning and delivering variants of three-dimensional conformal radiation therapy and intensity-modulated radiation therapy depends on the localization and digitization of anatomical structures. This task is currently accomplished by radiation oncologists that use a computer mouse on treatment planning workstations to draw outlines around CT reconstructions of structures. Although various tools have been provided to assist radiation oncologists with manual segmentation, the exercise remains a time-consuming bottleneck in the treatment planning procedure. In this talk we will present a semi-automatic technique for addressing this open problem. Our approach incorporates information about the target organ into classical segmentation algorithms. This information, which is given in terms of a 3D wire-frame representation of the organ, serves as an initial guess for the segmentation algorithm. Our method is based on solving nonlinear PDEs for the evolution of the curve that identifies the segmented organ with a velocity function that is taken as a piecewise Mumford-Shah functional. This is a joint work with Frederic Gibou, Carlos Cardenas, Pingyu Liu, and Arthur Boyer. |