A novel MLEM stopping criterion for unfolding neutron fluence spectra in radiation therapy

Characterization of a collimated neutron imager for low-rate fast neutron imaging

“…Poisson after inclusion of a noise factor and 'Monte Carlo agreement' [18] where iterations stop when the result has good agreement with Monte Carlo simulations. These types of stopping criteria depend on a priori knowledge about the data, which may not always be the case.…”

The impact of treatment parameter variation on secondary neutron spectra in high-energy electron beam radiotherapy

“…However, the conventional MLEM algorithm lacks an objective stopping criterion and introduces noise to the solution as it iterates. To reproducibly obtain noise-free spectra, we used a stopping criterion that was recently developed by our group [45], which is a modified version of the MLEM-STOP criterion of Ben Bouallègue et al [46] for PET image reconstruction. This criterion involves evaluating the following indicator function at each MLEM iteration k:…”

“…where I is the number of NNS moderator configurations, m i are the NNS measurements, and q k i are the MLEM-reconstructed measurements at a particular iteration, k. MLEM is terminated when J k ⩽ m 3×10 5 CPS . This threshold is based on the statistical properties of the NNS measurements, as described in detail in our previous publication [45].…”