MED In the literature: partially resolved

Optimal energy thresholds for material separation

In plain words

The detector sorts photons into a few energy bins. Where to place the bin edges for the best material separation at the lowest dose is unsolved in general.

Precise statement

Determine the energy threshold set that minimizes the variance of material-decomposition estimates for a given task, tube spectrum, patient thickness and dose, including the detector response with charge sharing and pileup. Compare against the Cramer-Rao bound (the minimum variance any unbiased estimator can reach) achievable with full spectral information. An answer is an optimization rule plus the efficiency loss from using 2 to 8 bins.

What would settle it

Derivation of the bound plus phantom measurement confirming the predicted variance for optimized thresholds.

Status in the literature

Information-theoretic treatments exist for idealized detectors; the realistic-response case stays open.