Combined tracking in the ALICE detector

A neural network based algorithm to perform track recognition in the ALICE Inner Tracking System (ITS) for high transverse momentum particles (p //t greater than 1 GeV/c) is presented. The model is based on the Denby Peterson scheme, with some original improvements which are necessary to cope with the very high track density expected in ALICE. The application is used in combination with the standard tracking procedure for track reconstruction in ALICE in order to increase the efficiency, especially for rapidly decaying particles. Results are shown for a test performed simulating some central Pb-Pb events at 5.5 ATeV in the center of mass system.