Background subtraction and peak search from threefold gamma event data

Abstract A method for subtracting background from triple-coincidence γ events is presented. In our data set it was used to remove 40% of the noise without affecting photopeaks with intensity of >18 counts. An example of performance of Ward's clustering algorithm applied to three-dimensional photopeak searching is also presented. Several standard clustering algorithms were found to be applicable only to background-subtracted data.