Machine learning and genetic optimization for particle tracking at high seeding densities

Abstract

Overview on current results in using machine learning techinques and genetic optimization for particle trackign at high seeding densities. We present a neural network based peak detection scheme which is then paired with a genetic optimization aproach for parameters of the iterative particle reconstruction. Synthetic as well as real world data is used to validate the approach and some preliminary results of incorporating such a scheme into a full Shake-the-Box evaluation and the achived improvements are shown