Method of fan-beam imaging calibration

We are proposing a simple and practical calibration method based on non-linear fitting of the measured point source sinogram coordinates to the geometrical model. We demonstrate that single set of point source sinogram coordinates from one acquisition does not uniquely define the calibration parameters. On the contrary, there are an infinite set of values for the calibration parameters and the source positions which provide the same sinogram values. In order to find calibration parameters by non-linear fitting, some additional constraints need to be applied to the optimization problem. More complicated phantoms are usually used to solve this problem. We suggest combining the acquisition data from two scans with a known difference of radii, which we can measure with high accuracy and include into the model. We show that this constraint provides a unique solution to the fitting problem. The results of calibration on phantom data are presented.