Identification of a region including metals in the projection data using an energy discriminating photon counting detector

Metal parts in a target object sometimes cause metal artifacts and beam hardening effects in an x-ray tomographic image. In this paper, we proposed a new method to extract a region of metals by finding a change of an energy spectrum in the projection data measured with an energy discriminating photon counting detector. The change in the energy spectrum was calculated using two parameters; RAI (relative attenuation index) and SDI (spectrum deformation index), that were defined to the projection data measured with different energy windows. And a scatter plot (SDI vs. RAI) was used for the extraction of pixels including metals in the projection data. The feasibility of our method was evaluated with Monte Carlo simulations with a numerical phantom and experiments with two human teeth phantoms. The results of the simulations and experiments showed that pixels including metals and those without metals were separated from each other, making it possible to easily separate regions including metals in the projection data. The accuracy of extracting metal parts depended on its atomic number, however, pixels including metals with a large atomic number such as copper and silver could be easily separated.