Dynamic tomography with curve evolution methods

In this paper, we propose a curve evolution method for dynamic tomography. We jointly estimate both object boundaries and intensity dynamics using all the data. The image intensity is assumed piecewise constant and the activity in each region is modeled as a parameterized dynamic process. The boundaries of objects are modeled as a series of curves. The boundaries and dynamics are estimated jointly through a curve evolution process. This curve evolution process is implemented with level set methods, which can handle topology changes easily.