Determining Noise Distribution in Computed Tomography - A Simple Phantom Based Approach

In Computed Tomography (CT) the acquired data is always influenced by noise. This is due to the fact that i) the creation and interaction of X-rays is a statistical process and ii) the detector electronics have some inherent noise as well. Assuming the presence of monochromatic radiation, the values of the measured intensity values are Poisson distributed. Using the central limit theorem the noise distribution turns into Gaussian noise at higher dose levels. In this work, the noise distribution is measured using a simple phantom and compared to a simple noise model.