28 June 2021 to 2 July 2021
Europe/Vienna timezone

Automatic Classification of Particulate Radionuclide Spectra

1 Jul 2021, 09:00


e-Poster T3.5 - Data Analysis Algorithms T3.5 e-poster session


Mr Tryggvi Edwald (Independent Researcher, Vienna, Austria)


This poster describes progress in 2020 of classifying IMS particulate radionuclide spectra into “likely normal” (Cat. 1,2) and “requires scrutiny” (Cat.3,4,5).The aim is to assist IDC radionuclide analysts by reducing their workload of “normal” spectra, which are the majority of IMS spectra.
The method described here is entirely ignorant of radionuclide science. The point of these efforts was to find how close to a correct classification one could get without using any radionuclide science. Accuracy of over 94% has been achieved. It is likely that injecting radionuclide science into the method will improve the results.

Primary author

Mr Tryggvi Edwald (Independent Researcher, Vienna, Austria)

Presentation materials