Adaptive algorithms of position and energy reconstruction
by
Andrei Morozov, Vladimir Solovov(LIP-Coimbra)
→
UTC
LIP _ Lisbon
LIP _ Lisbon
Description
Scintillation camera, first developed by Anger in 1957, is still widely used in medical imaging as well as in experimental physics, for example in astrophysics and neutron detection. The position reconstruction is typically performed by the Centre-of-Gravity algorithm (dating back to original Anger's work) followed by application of some kind of a correction map to compensate for the image distortions. While relatively simple and robust, this method has two serious drawbacks: (i) it does not work well over the whole detector’s area and (ii) it requires frequent recalibrations to compensate for the drift of the PMT gains.
Statistical reconstruction algorithms (such as the Maximum Likelihood algorithm) can produce very accurate reconstructions; however, all these algorithms require detailed information on the spatial dependence of the PMT response – the so called Light Response Function (LRF), which is usually not trivial to obtain.
In this talk we will discuss the adaptive algorithms developed in LIP which allow to reconstruct the LRFs and relative PMT gains from a flood illumination data recorded with the detector, a measurement which is typically simple to perform in any practical application. We will show examples of successful application of these algorithms for dark matter detectors (ZEPLIN and LUX collaborations) and thermal neutron detectors (FP7-NMI3 collaboration). A publicly-available simulation package ANTS (Anger-type Neutron detectors: Toolkit for Simulations) developed in LIP, which was recently extended for processing of experimental data using the adaptive algorithms will be discussed. We will also present our plans for the future work, namely extending applicability of these adaptive techniques to the area of medical imaging.