University PhD Dissertation Defense<br><br>Title: <i>Photon Counting Detector Cross Talk Model and Optimizing Information in Spectral CT</i> <br>Speaker: Paurakh Rajbhandary, Advisor: Nobert J. Pelc<br>Date: November 17, 2017<br>Time: 1:00pm (refreshments at 12:45pm)<br>Location: Packard 101<br><br>Abstract: <br>Spectral CT can provide additional energy specific information that can be diagnostically useful and could provide supplementary information, such as tissue specificity, to conventional CT systems. Currently, dual source, fast kVp-switching, dual filter and dual layer detector systems, all of which use energy integrating detectors (EIDs), are available clinically to enable spectral CT. Energy discriminating photon counting detectors (PCDs) have been shown in research to extract more information than EIDs and from a single exposure. However, charge sharing, scatter and fluorescence events in a PCD can result in multiple counting of a single incident photon in neighboring pixels. This causes energy spectrum distortion and correlation of data across energy bins in neighboring pixels (spatio-energy correlation). Spatial correlation in conventional linear, space-invariant imaging system can be usefully characterized by the frequency dependent detective quantum efficiency DQE(f). Estimate of DQE(f) for PCDs and spectral imaging is complicated due to the energy dimension. In this work, we present an analysis of DQE(f) of CdTe PCDs using a “brute-force” method starting from an analytic computation of spatio-energetic cross talk based on pre-computed functions from Monte Carlo simulation. In this talk, I will describe methods for modeling detector cross-talk and also explore the effect of such phenomena as well as electronic noise in PCD performance.