A correction framework for non-ideal source switching in dual energy CT scanner

Dual energy CT requires two measurements of object attenuation at two different energies. Two measurements can be obtained by switching between two sources operating at two different spectra. Non-ideal source switching causes data mixing which produces image artifacts. We propose a framework to correct data to prevent such image artifacts using a recursive view mixing model. A nonlinear optimization approach estimates the model parameters using data from a calibration procedure based on a symmetry metric. Results from a commercial dual energy CT scanner demonstrate the effectiveness of the proposed approach.