The team will conduct research on fast image reconstruction using two novel convergent maximum-likelihood (ML) reconstruction methods employing iterative algorithms based upon ordered-subsets principles. These algorithms, unlike others, are guaranteed to converge regardless of the presence of noise, and regardless of whether any subset balance conditions are met. They will investigate techniques for detecting the change of recorded gamma rays in spatial, spectral, and timing domains. The energy spectra will then be analyzed in order to identify the specific isotope, by using wavelet analysis, feature extraction and reconstruction, and pattern recognition to compare theoretical (from a library) and measured patterns. The specific isotope will then be identified by a decision-making strategy. The project will make contributions to science and knowledge in radiation detection, signal processing, identification, and locating of highly enriched uranium, plutonium, and other radioactive isotopes that might be used for "dirty bombs".
