Software-based charge sharing correction for spectroscopic x-ray detectors

Spectroscopic x-ray detectors, which are currently under development, may have the potential to revolutionize diagnostic x-ray imaging through their ability to perform single-shot, multi-energy imaging. However, the sharing of charge between neighbouring detector elements results in degraded image quality. The goal of this project was to develop a computer algorithm to correct for the effects of charge sharing. The 3 primary objectives were: (1) to develop a model of the system response of spectroscopic x-ray detectors which accounts for the spatio-energetic effects of charge sharing, (2) to develop a charge sharing correction algorithm, based on the maximum likelihood expectation-maximization (MLEM) algorithm which incorporates the spatio-energetic system response model and (3) to assess the accuracy of the charge sharing correction algorithm by applying it to simulated x-ray images. The MLEM algorithm was able to correct for the spectral distortion resulting from charge sharing with reasonable accuracy. Its performance was somewhat erratic due to the non-linearity of the algorithm, but was found to perform better when applied high-energy x-ray spectra.